Gap colonization in grasslands of the Northern Great Plains
by Elibeth Mary Payson
A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in
Biological Sciences
Montana State University
© Copyright by Elibeth Mary Payson (1996)
Abstract:
The effects of gap size on colonizer success was examined in two intermountain grasslands during two
years. Eight species, four grasses and four forbs with a range of seed size and rooting systems was
considered. Gaps ranging from 1 cm2 to 1 m2 were created in two sites. Measurements were taken on
plant emergence, survival, biomass, and height on each plant. Environmental factors are examined to
help explain the results.
Results indicate that gap size affects plant emergence, survival, biomass, and height. Emergence is
little affected by gap size. Survival, especially in drier years, tends to increase with increasing gap size
and significantly so when disturbances reach 1 m2. Biomass, an indicator of plant growth, tends to
increase with gap size and this increase becomes significant with gap sizes of 1 m2. Another index of
plant growth, height, shows similar behavior.
Survival and growth are affected by physical and biological factors. Thus, germination tended to be
greater in a moister site. Available water plays an important role in both grasslands.
Key Words: gap colonization, competition, grassland GAP COLONIZATION IN GRASSLANDS OF
THE NORTHERN GREAT PLAINS
by
Elizabeth Mary Payson
A thesis submitted in partial fulfillment
o f the requirements for the degree
of
Master o f Science
in
Biological Sciences
MONTANA STATE UNIVERSITY
Bozeman, MT
April, 1996
a/3'1?
a
APPRO V A L
o f a thesis submitted by
Elizabeth Mary Payson
This thesis has been read by each member o f the thesis committee and has been
found to be satisfactory regarding content, English usage, format, citations, bibliographic
style, and consistency, and is ready for submission to the College o f Graduate Studies.
Theodore W. Weaver III
I
Chairperson, Graduate Committee
C
Date
Approved for the Department of Biology
Ernest R. Vyse
/7 ^ /0
Department Head
Date
Approved for the College o f Graduate Studies
Dat
in
STA TEM EN T O F PE R M ISSIO N TO USE
In presenting this thesis in partial fulfillment o f the requirements for a master's
degree at Montana State University—Bozeman, I agree that the Library shall make it
available to borrowers under the rules o f the Library.
If I have indicated my intention to copyright this thesis by including a copyright
notice page, copying is allowable only for scholarly purposes, consistent with "fair use" as
prescribed in the U.S. Copyright Law. Requests for permission for extended quotation
from or reproduction o f this thesis in whole or in parts may be granted only by the
copyright holder.
A CK N O W LED G EM EN TS
I would like to expresss my appreciation and thanks to the following people and
departments for their assistance during the course o f my studies:
To Professor T. Weaver for being so generous with his time, so incredibly patient,
and interested in people's ideas. Dr. Weaver has a genuine enthusiasm for scientific
questions which is infectuous, and I particularly appreciate his liberal arts approach to
academia.
To my committee members, Dr. Bruce Maxwell, Dr. Rich Stout, and Dr. Bret
Olsen, for their support in preparing this thesis. Especially to Dr. Bruce Maxwell, who
assisted in the planning and analysis o f this work.
To Sigma Xi, the Scientific Research Society for a Grant-in-aid-of-Research to
help support this study. Also to the Biology Department for financial assistance provided
in the form o f a Graduate Teaching Assistantship.
Tb my parents for providing me with the opportunity to pursue my interests, for
instilling in me the curiosity and skills to follow it, and for their constant support.
vi
TABLE O F CONTENTS
Page
VITA............................................................................................................................ h,
ACKNOWLEDGEMENTS...................................................................................... v
TABLE OF CONTENTS.......................................................................................... vi
LIST OF TABLES.................................................................................................... viii
LIST OF FIGURES......:............................................................................................ xii
ABSTRACT................................................................................................................. xiii
INTRODUCTION...................................................................................................... I
Background.............................................................. ........................................... I
Problem Statement......... .................................................................................... I
Literature Review................................................................................................ 2
M ETHODS...................................................................................................................6
Site Description................................................................................................... 6
Data Collection................................................................................................... 6
Analysis................................................................................................................. 9
RESULTS.....................................................................................................
Introduction..........................................................................................................
Plant Performance................................................... .:.......................... ...............
Emergence Response to gap size, site, species, and year..................
Survival Response to gap size, site, species, and year........................
Biomass Response to gap size, site, species, and year.......................
1993 Biomass Regressions........,...........................
1993 Biomass Anovas...............................................................
1994 Biomass Regressions.......................................................
Forb v. Grass Regressions.........................................................
1994 Biomass Anovas...............................................................
1993 v. 1994 Biomass Regressions.........................................
1993 v. 1994 Biomass Anovas..;.............................................
14
14
14
14
15
16
17
17
17
17
18
19
19
vii
TABLE O F C Q N T E N T S -Continued
Height Response to gap size, site, species, and year............................ 20
1993 Height Regressions............................................................ 20
1993 Height Anovas.................................................................. 20
1994 Height Regressions............................................................ 20
1994 Height Anovas.................................................................. 21
Forb v. Grass Regressions.......................................................... 22
1993 v. 1994 Height Regressions.............................................. 22
1993 v. 1994 Height Anovas......................................................22
Seed Production...................................................................................... 23
Environmental Characteristics o f the Sites..............
23
Precipitation..................................................
23
Grasshopper Herbivory..............
24
Light Intensity....................................................................
24
Soil M oisture...................................
25
DISCUSSION................................................................................................................26
Emergence.............................................................................................................. 26
Survival................................................................................................................. 27
Gap Size................................................................................................... 27
Site........................................................................................................... 28
Year..................... •................................................................................... 28
Species....................................................................................
29
Year and Species Interaction............................................................... 29
Biomass................................................................................................................ 29
1993 Gap Size........................................................................................ 29
1994 Gap Size......................................................
30
1993 Site...........................................................
32
1994 Site............. :.................................................................................. 32
Grass v. Forb......-..................................
33
Seed Size................................................................................................. 34
Year............................................................................................................ 33
Height...................................................................................................................... 34
Gap...............................
34
Site and Y ear.......................................................................................... 35
Seed Production.................................................................................................. 35
CONCLUSION............................................................................................................. 36
LITERATURE CITED............................................................................................... 38
APPENDIX................................................................................................................. 43
via
LIST O F TABLES
Table
Page
1. Characteristics o f Plants Treated............................................................................... 44
2. Emergence (%) o f all Species at Five Gap Sizes..................................................... 45
3. Emergence (%) and Survival (%) at Each Date and Site................. ...................... 46
4. Total emergence (%) for Each Plant......................................................................... 47
5. Effect o f Gap Size, Site, and Species on Plant Survival (%), 1993...................... 48
6. Analysis o f Variance on W heat and Knapweed Survival Data,1993..................... 48
I. Analysis o f Variance on Survival Data from 1994.................................
49
8. Effect o f Gap Size, Site, and Species on Plant Survival (%),1994........................ 50
9. Analysis o f Variance on Survival Data, 1993 and 1994......................................... 51
10. Effect o f Site and Year on Survival (%) o f Knapweed and Wheat
1993 and 1994................. '................................................................ ........................ 51
II. Relationship o f Plant Weight (g) to Site, Year, and Hole Size........................... 52
12. Analysis o f Variance on Wheat and Knapweed Weight from Both
Sites, 1993....................................
53
13. 1993 Plant Weights (g) at Both Sites, 1993............................. ............................. 53
14. Analysis o f Variance, Wheat and Knapweed" Weight, Blue Grama
Site, 1993................................................................................................................... 54
15. Analysis o f Variance, Wheat and Knapweed Weight Idaho Fescue
Site, 1993.............
54
16. Wheat and Knapweed Weights (g) at the Blue Grama Site, 1 9 9 3 ..................... 55
17. 1993 W heat and Knapweed Weight (g) at the Idaho Fescue Site, 1993............ 55
18. Grass and Forb Height and Weight in Relationship to Gap Size........................ 56
ix
L IST O F TA BLES-Continued
Table
Page
19. Analysis of Variance Plant Weights (Log Transformed) at Both
Sites, 1994..............................................................................................................• 57
20. Effect o f Gap Size on Plant Weights (g), Both Sites, 1994.................................. 57
11
21. Plant Weights (g) at the Blue Grama Site, 1994.................................................. 58
22. Analysis o f Variance on Plant Weights, Blue Grama Site, 1994 ......................... 58
23. Plant Weights (g) at the Idaho Fescue Site, 1994................................................ 59
24. Analysis o f Variance on Plant Weights (Log Transformed) at the Idaho
Fescue Site, 1994..................................................................................................... 59
25. Analysis o f Variance on Wheat and Knapweed Weight at Both Sites,
1993 and 1994........................................................................................................... 60
26. Effect o f Gap Size on Wheat and Knapweed Weights (g) at Both Sites,
1993 and 1994........................................................................................................... 60
27. Analysis o f Variance on Plant Weights at the Blue Grama Site,
1993 and 1994............................................................................................................ 61
28. Wheat and Knapweed Weights (g) at the Blue Grama Site,
1993 and 1994................................................................... ....................................... 61
29. Analysis o f Variance on Weights at the Idaho Fescue Site,
1993 and 1994.......................................................................................................... 62
30. Wheat and Knapweed Weights (g) at the Idaho Fescue Site,
1993 and 1994........................................................................................................... 62
31. Relationship o f Plant Height to Site, Year, and Gap Size......................................63
32. Analysis o f Variance on Wheat and Knapweed Heights at Both Sites,
1993.................................................... ' ......................................................................
64
Y
■I
L IST O F T A B L E S -Continued
Table
Page
33. Effects o f Gap Size on Wheat and Knapweed Heights (cm) at
Both Sites, 1993....................................................................................................... 64
34. Analysis o f Variance on Wheat and Knapweed Heights at the
Blue Grama Site,. 1993............................................................................................ 65
3 5. Analysis o f Variance on Wheat and Knapweed Heights at the
Idaho Fescue Site, 1993.........................................................................................
65
36. Wheat and Knapweed Heights (cm) at the Blue Grama Site, 1 9 9 3 .....................66
37. Wheat and Knapweed Heights (cm) at the Idaho Fescue Site, 1993................. 66
38. Plants Heights (cm) at Both Sites, 1994........................................................
67
39. Plants Heights (cm) at the Blue Grama Site, 1994...........
68
40. Plants Heights (cm) at the Idaho Fescue Site, 1994..........................................
69
4 1. Analysis o f Variance on Wheat and Knapweed Heights at Both Sites,
1994........................................................................................................................... 70
42. Analysis o f Variance on Wheat and Knapweed Heights at the Blue
Grama Site, 1994..............................
70
43. Analysis o f Variance on Wheat and Knapweed Heights at the Idaho
Fescue Site, 1994...................................................................................................... 71
44. Effect o f Gap Size on Wheat and Knapweed Heights (cm) at
Both Sites, 1993 and 1 9 9 4 ....................................................................................... 72
45. Analysis o f Variance on Wheat and Knapweed Heights at Both
Sites, 1993 and 1994.................................................................
72
46. Analysis o f Variance on Wheat and Knapweed Heights at the Blue
Grama Site, 1993 and 1994............................ ........................................................ 73
47. Analysis o f Variance on Wheat and Knapweed Heights at the Idaho
Fescue Site, 1993 and 1994..................................................................................... 73
L IST O F TABLES—Continued
Table
Page
48. Effect o f Gap Size onWheat and Knapweed Heights (cm) at the
Blue Grama Site, 1993 and 1 9 9 4 .......................................................................... 74
49. Effect o f Gap Size on Wheat and Knapweed. Heights (cm) at the
Idaho Fescue Site, 1993 and 1994......................................................................... 74
50. Precipitation Data at the Blue Grama Site, compared to Long Term.................. 75
5 1. Precipitation Data at the Idaho Fescue Site, compared to Long Term............ 76
52. Percentages o f Plants Eaten By Grasshoppers at Both Sites, 1993
and 1994...................................................................................................................
77
53. Signed Rank Sum Test Results for Grasshopper Control............................
77
54. Light Intensity (Pages) in Gaps............................................................................... 77
55. Soil Moisture (%) at Both Sites, 1994..........................
78
56. Emergence (%) o f Eight Species at Both Sites, 1993 and 1994......................... 79
57. Survival (%) o f Eight Species at Both Sites, 1993 and 1994...........................
80
58. Knapweed and Wheat Heights and Weights at the Blue Grama
Site, 1993................................................................................................................... 81
59. Knapweed and Wheat Heights and Weights at the Idaho Fescue
Site, 1993.................................................................................................................. 82
60. Grass Heights and Weights at the Blue Grama Site, 1994................................... 83
61. Grass Heights and Weights at the Idaho Fescue Site, 1994.............................. 84
62. Forb Heights and Weights at the Blue Grama Site, 1994.................................... 85
63. Forb Heights and Weights at the Idaho Fescue Site, 1994
86
xii
LIST O F FIG U RES
Figure
Page
1. Effect og Gap
Size on Forb Weight (Log Scale)........................................... 87
2. Effect og Gap
Size on Grass Weight (Log Scale)........................................... 87
3. Effect og Gap
Size on Forb Height (Log Scale)............................................ 88
4. Effect og Gap
Size on Grass Height (Log Scale)........................................... 88
5. Clover Weight at Both Sites....................................................................................... 89
6. Sunflower Weight at Both Sites................................................................................. 90
7. Crested Wheatgrass Weight at Both Sites............................................................... 91
8. Cheatgrass Weight at Both Sites............................................................................... 92
9. Com Weight at Both Sites......................................................................................... 93
10. Knapweed Weight at Both Sites, 1993................................................................... 94
11. Knapweed Weight at Both Sites, 1994................................................................... 95
12. Wheat Weight at Both Sites, 1993......................................................................... 96
13. Wheat Weight at Both Sites, 1994
97
xiii
ABSTRA CT
The effects o f gap size on colonizer success was examined in two intermountain
grasslands during two years. Eight species, four grasses and four forbs with a range o f
seed size and rooting systems was considered. Gaps ranging from I cm2 to I m2 were
created in two sites. Measurements were taken on plant emergence, survival, biomass,
and height on each plant. Environmental factors are examined to help explain the results.
Results indicate that gap size affects plant emergence, survival, biomass, and
height. Emergence is little affected by gap size. Survival, especially in drier years, tends
to increase with increasing gap size and significantly so when disturbances reach I m2.
Biomass, an indicator o f plant growth, tends to increase with gap size and this increase
becomes significant with gap sizes o f I m2. Another index o f plant growth, height, shows
similar behavior.
Survival and growth are affected by physical and biological factors. Thus,
germination tended to be greater in a moister site. Available water plays an important role
in both grasslands.
Key Words: gap colonization, competition, grassland
I
• IN T RO D U C TIO N
Background
Due to intense competition, plants rarely establish in undisturbed vegetation.
Disturbance creates gaps which allow change in grasslands through lateral growth of
dominants or seeds from mobile natives, mobile exotics (both opportunists), or less
mobile natives. We hypothesize that the likelihood of establishment o f a plant from a
seed increases as the size o f disturbance increases, that is, as the probability of
reinvasion by competing roots, stolons, or rhizomes diminishes.
Problem Statement
Natural disturbances in grasslands may range in size from worm holes (I cm2),
hoof prints (10 cm2), gopher mounds (100 to 1000 cm2), and elk or bison wallows (I
m2). The object o f this paper was to compare the establishment success of diverse
plants in these five hole sizes in two very different grasslands and two very different
years. W e tested four hypotheses: I) Germination does not vary with gap size because
the limiting resource, water, is abundant in the spring, 2) Establishment and growth is
better in larger disturbance sizes. 3) Establishment and growth are better in molster
grasslands and moister years, 4.) Forbs establish more easily than grasses in
grasslands because their taproots compete less with the diffuse roots o f grasses.
The study was conducted in two grasslands, one dry and one wet, over a dry
year and a moist year. Success o f plants representing varying seed size, rooting
2
strategy, colonizing strategy, and growth form was compared. Our results shed light on
community structure, illuminating the role that gaps play in natural community
dynamics. W e simultaneously show how weeds respond to natural disturbances.
Literature Review
Gaps are places in a plant community where either the shoot or root canopy
have been disrupted. These disturbances are normal in grassland ecosystems, and
provide space which plants may colonize. Tilman (1982) offers two ways to think o f
this space. The first is as an open space devoid o f other organisms. The second,
which he argues as the more pertinent, is as the sum of the resources which it offers.
Lieberman and Leiberman (1989) echo this in their article "Forests are not just swiss
cheese." Commonly gaps, or disturbances, are considered as a break in an otherwise
continuous vegetative cover. Lieberman and Lieberman (1989) argue that this is the
wrong supposition, and that to assess the effect o f a disturbance, the entire community
including gaps, must be included in a comparison of a particular gap to the fabric o f a
forest. It is beneficial, therefore, to compare grasslands o f different densities, because
they do not behave similarly. The importance o f gaps and patchy landscapes have been
emphasized by numerous authors (Tilman, 1982; Pickett and W hite, 1985a;
McConnaughay and Bazzaz, 1987).,
Gaps play a central role in community structure, diversity, development, and
succession in grasslands. Once considered to be a stable community in which
disturbance was infrequent (Clements and Shelford, 1939) researchers have come to
3
recognize the importance o f disturbances caused by animals, fire, and erosion (Arno
and Gruel, 1983; Daubenmire 1978; Umbanhower, 1991). All three authors assert that
the occurrence o f fire in grasslands prevents encroachment by trees into the grassland.
Disturbances caused by animals to plants or the soil are addressed by many authors as
well. (McCaunnaghay and Bazzaz, 1990; Coffin and Lauenroth, 1989; Coffin and
Lauenroth, 1990; Coffin and Lauenroth, 1994; Fenner 1980). The nonequilibrium
approach, the idea that plant communities are constantly subjected to disturbances and
different areas are in vaious stages of recovery, is clear (Reice, 1994). Small scale
gaps influence succession by returning parts o f a site to lower serai stages, and allowing
for colonizer species to invade (Platt, 1975). This creates a mosaic o f patches in
various levels o f succession, and increases diversity.
We will compare growth of first year colonizers over an exponential range of
gap sizes which has not been addressed in previous literature. Plants colonize
disturbances ranging from less than Icm 2 to whole agricultural fields. Past studies
examining disturbances are summarized below. Umbanhower studied natural
colonization of 7.5 dm artificially created mounds in northern prairies (1991)
Heterogeneity has been examined at a scale o f gaps less than 30 cm (Hook and
Lauenroth, 1994a and 1994b; Hook, Burke, and Lauenroth, 1991; Reader, et al,
1994).. In the tail-grass prairie, Platts (1975) Studied the effect o f badger mounds on
succession. The effect o f size and frequency o f disturbance has been examined by
Campbell, Grime, and Mackey (1991), and Coffin and Lauenroth (1988).
Monospecific interactions as well as interactions between two species have been
4
examined extensively (Tilman, 1982, 1987; Ang, et al, 1995; Collins and Rhodes,
1994). Neighborhood interactions have received ample treatment (Harper, 1977;
Aguilera and Lauenroth, 1993a and 1993b; Fowler, 1988). Patches have also been
examined at the scale of landscape level succession (Tilman, 1988; Reice, 1994; Coffin
and Lauenroth, 1990; Forman and Godron, 1981).
Colonizer species invade disturbances. Their success depends upon attributes
which enable them to germinate, survive, and reproduce in a gap, Varying gap sizes
select for different life histories of invading annuals (McCbnnaughay and Bazzaz,
1987).
Several factors influence colonization success in grassland gaps. These are
seed size, germination strategy, seedling growth, and regenerative and rooting strategy.
The study o f colonizer strategy must address individual life stages as success during
each influences overall plant performance. Grime states that a p lan t's success depends
upon a combination of these attributes, and that characteristics which are beneficial in
one situation may not be beneficial in another (Grime 1979). Plants representing a
range o f these habits were chosen for this study.
Seed characteristics are important determinants of colonizer strategy. Fenner
(1980) studied seed characteristics in relation to plant succession. He found that seed
weight increases from early serai to late. He also found that light sensitivity was
important in gap colonization. Fenner (1978) asserted that gap colonizers germinate
when they detect wide fluctuations in temperature. Temperature extremes are greater
in gaps than in canopies. Seeds can detect the presence of openings because of these
temperature extremes. In comparing seedling emergence times, the earlier a seed
5
germinated, the more successful a competitor it was (Fowler, 1988, Harper, 1977).
Early seedling growth is also important to colonization success. Large size, and
the ability to dominate competitors bestow a great advantage to a plant (Grime 1979) .
One plant may dominate another by preempting soil resources such as nutrients or
water, and intercepting light before it can reach a competitor. Harper (1977) described
how the emergence date and growth rate enable a plant to overwhelm competitors by
changing a site's physical and chemical characteristics. It is to a colonizing plant's
advantage to have rapid seedling growth.
For all lifeforms, the ability to reproduce has been recognized since Darwin as
the true measure o f an individual’s success. Regenerative strategy takes on a number
o f forms in colonizers. Annuals are common colonizers. They invest energy in rapid
growth and seed production. The annual strategy is completely dependent on seed
production within one growing season (Grime, 1979). Biennials allow themselves a
season to accumulate resources before producing seed, and perennials rely on long term
resource allocation and repeated seed production for success.
6
M ETH O D S
Site Description
The effect o f gap size on plant performance and conditions that may influence that
success— site conditions, yearly precipitation, and colonizer species—were determined.
Establishment was compared in two grassland communities. The first, was a Festuca
idahoemis/Agropyron spicatum (Idaho fescue/ bluebunch wheatgrass) community, was a
moist (89 cm/yr p p t), densely populated grassland with productive soils (Mueggler and
Stewart, 1980; Weaver, 1979). This site was located in a pristine field in Bozeman, MT
(T2S, R6E, S 21, N E 1/4, NE 1/4). The second was in a Stipa comata/Bouteloua
gracilis (needle-and-thread grass/ blue grama) community; with drier (73 cm/yr ppt), finer,
less productive soils (Mueggler and Stewart, 1980). This site is three miles west of
Logan, MT, near Interstate 90 (T2N, R2E, S26, SE 1/4, SE 1/4). The Idaho fescue/
bluebunch wheatgrass site and the needle-and thread grass/ blue grama site will be referred
to as the Idaho fescue (Feid) and blue grama (Bogr) sites respectively.
Data Collection
We tested the effect o f gap size on colonization success. To do this, we made
gaps in two grasslands over two growing seasons. Study blocks in each grassland were
chosen for uniform terrain and visual homogeneity. Small plots ranging in size from Icm2
through 10, 100,1000, to 10,000 cm2 were disturbed. In each disturbance size, soil was
removed to, a depth o f IOjxn, sieved with a I cm mesh screen and mixed in a central
7
location, and returned to the disturbance. Roots, rhizomes and rocks were removed in the
sieve. Uniformity was insured by the use o f I cm2 and 10 cm2 cores for the smaller holes,
and careful measurement and excavation with trowels and shovels for larger holes. Soil
was removed from all holes, mixed, and returned to insure homogeneity in soil texture,
nutrients, and treatment.
Many replications were used to support statistical tests. In 1993, each replication
included five I cm2, three 10 cm2, one 100 cm2, and one 1000 cm2 gap. Also, one 10,000
cm2 disturbance was created for each plant. Five replications were made for each plant at
each site. Thus a total o f twenty-five I cm2, fifteen 10 cm2, five 100 cm2, five 1000 cm2
and one 10,000 cm2 disturbances were studied at each site. In the second year more 100
and 1000 cm2 holes were created. The 10,000 cm2 disturbances were divided into
quarters, and one plant was grown in each quarter. This was done to minimize damage to
the site, and create replications at the largest disturbance size. In 1994, five I cm2, three
10 cm2, two 100 cm2, and two 1000 cm2 gaps were created. Again, five replications were
planted for each species. Thus, establishment was observed in twenty-five I cm2, fifteen
10 cm2, ten 100 cm2, and ten 1000 cm2 gaps, and four replications o f the I m2 gap for
each plant in 1994.
Growth was observed between May and August. In 1993, the Idaho fescue sites
were planted on 22 May, and the blue grama site was planted on 29 May. Emergence and
height measurements were recorded every two weeks. Plants were harvested on 13
August at the Idaho fescue site, and 20 August at the blue grama site. Plant height and
weight were documented at harvest. In 1994, seeds were planted earlier to take
8
advantage o f early season precipitation. Seeds for most plants were planted at the blue
grama site on 7 April and at the Idaho fescue site on 14 April. Corn and sunflower were
planted on 10 May at both sites. Emergence, height, stem diameter, and volume were
recorded every two weeks throughout the summer. Plants were harvested at the blue
grama site on 12 August, and at the Idaho fescue site on 23 August. When harvested,
plant height, canopy width, stem diameter, and weight were measured.
Eight species were studied. In 1993 knapweed (Centaurea maculosa) and wheat
(Triticum aestivum cultivar Winridge) were planted. In 1994, we wanted to contrast
performance o f a greater variety o f plant species. We planted four grasses: wheat
(Triticum aestivum cultivar Winridge), cheatgrass (Bromus tectorum), crested wheatgrass
(Agropyron cristatum), and com (Zea mays cultivar Silver bullet), and four forbs: spotted
knapweed (Centaurea maculosa ), yellow sweet clover {Melilotus officinalis), shepherd's
purse (Capsella bursa-pastoris), and sunflower (Helianthus annuus cultivar Giant
Greystripe). Nomenclature follows Hitchcock and Cronquist (1973) and Gould and Shaw
(1983). These plants were chosen for their range o f seed size, rooting strategy, lifeform,
and colonization strategy. Table I summarizes the species' characteristics.
Grasshoppers destroyed many plants during the 1993 field season. To minimize
this, we treated the sites with Sevin insecticide on bran bait every two weeks in 1993.
Control was inadequate, so in 1994, sites were treated each week.
Establishment and growth differences across gap size might be due to differences in light, temperature, or resource (water and nutrient) availability. These characteristics
were therefore measured across gap sizes. Light intensity was compared across gap sizes
9
using the Friend (1961) method. Twenty-five sheets o f blueprint paper, 2.5 cm x 2.5 cm,
were assembled into booklets and wrapped in black construction paper envelopes. A 0.4
cm hole was punched in the envelope with a hole punch, so light could penetrate only in
this spot. These envelopes were placed in 4 cm petri dish to exclude rainfall. Five o f
these were placed in each disturbance size for 24 hours. After removal, the papers were
kept together in their black construction paper envelopes and developed with ammonium
hydroxide fumes for 15 minutes. Light exposed areas become yellow and unexposed areas
are blue. The number o f exposed sheets were counted as an index o f fight intensity. The
fight levels in the different disturbance sizes were compared with analysis o f variance.
Soil moisture was measured on three dates throughout the 1994 field season
(Idaho fescue site: 5/15/94, 7/22/94, and 9/8/94; blue grama site: 5/24/94, 7/15/94, and
9/8/94). Samples were collected to a depth.of 10 cm from each gap size and placed in a
fined paper bag. Three bags were taken from each gap size on each date, except June 15,
in Bozeman, when only two bags per gap were collected. Moist weights were recorded,
then samples were dried in an oven at 60 C for 2 weeks, then dry weights were taken.
Soil moisture % was calculated as grams o f water divided by grams o f soil.
Temperature was measured in each gap size onfour dates. We do not present this data
because effects o f time o f day confounded effects o f gap size.
Analysis
Each year's data were analyzed separately. In addition a pooled analysis o f
knapweed and wheat over both years was-performed. The 1993 analysis shows how
10
wheat and knapweed respond to a wet year, 1994 data show how they respond to a dry
year, and simultaneous analysis illuminates the difference between an exceptionally wet
and a dry summer. 1994 analysis o f other plants permits comparison o f gap size effects on
different species o f plants in a dry year. Effect o f disturbance size, grassland type, and
lifeform was determined for individual species and for all plants together.
Data sets were first tested for normal distributions and equal variance using the
Wilks-Shapiro test (Winer, 1962) and Hartley's test respectively (Neter, Wasserman, and
Kutner, 1990). Additionally, residuals were graphed against normal scores, gap size, and
fitted values to visually confirm normality and equal variance.
Emergence was related to gap size, emergence date, and species on plant
emergence percentages using Kruskall-Wallace (Neter, Wasserman, and Kutner, 1990)
and Mann-Whitney tests (Neter, Wasserman, and Kutner, 1990). Differences in
emergence across disturbance size, dates, and species were analyzed as categorical
variables using Kruskall-Wallace analysis o f variance' on ranks and Dunn's test for multiple
pairwise comparisons (Winer, 1962). This non-parametric test was used because these
data were not normally distributed. Since there was no reason to believe that a trend
would exist between disturbance sizes, dates, or species, analysis o f variance was
appropriate. Dunn's test for pairwise comparisons was used for all Kruskall Wallace tests,
since it can be used with non-parametric tests and on unequal sample sizes. MannWhitney tests (Neter, Wasserman, and Kutner, 1990) were used to compare plant
treatments, years, and sites in 1993.
Plant survival was related to gap size, site, year, and species using analysis o f
11
variance and Tukey's pairwise comparisons (Neter, Wasserman, and Kutner, 1990). All
possible combinations o f factors and their interactions were tried, and this was the best
model. This test was conducted on 1993 and 1994 data separately, and on the 1993-1994
knapweed and wheat data together with the addition o f the year factor included.
Height and weight responses to gap size, site, species, and year were treated as
categorical variables using analysis o f variance and Tukey's pairwise comparisons. The
Box-Cox procedure was used to find the best transformation o f the data for analysis o f
variance (Neter, Wasserman, and Kutner, 1990). Based on this, the log transformation
was used for anovas. The appropriate Anova model used the response variable, log
(weight or height), and the factors site, gap size, species, and year. Before analysis, the
mean o f responses for each species at each gap size was calculated and used in anova.
This created equal sample sizes, one o f the assumptions for Tukey's pairwise comparisons.
Differences in biomass and height across gap size, site, year, species, and habit
were also treated as continuous variables to which regression analysis was applied. Single
species weights and heights were regressed against site, log(gap size), and log(gap size)2
for the 1993 and 1994 data separately. To compare wheat and knapweed growth in the
two years, each plant's weight and height were regressed against site, year, log(gap size),
and log(gap size)2. Examination o f residual distribution indicated that these same factors
provided the best fit line for most other species. In some cases, the best fit line did not
include all o f these factors (year was only used in comparing knapweed and wheat
responses both years), or the log o f plant height or weight yielded a better expression o f
the data (less the 50% o f height data are log transformed). This is indicated in the results.
12
The Durbin-Watson test was used to test for autocorrelation (Neter, Wasseman, and
Kutner3 1990)
Performance o f grasses and forbs was also contrasted with regression; Partial F
tests were used to compare the performance o f grasses versus forbs. This analysis was
conducted on the 1994 height and weight data only. The log o f plant weight or height
was regressed against the log o f gap size, log o f disturbance size squared, a code o f I or 0
for forb or grass respectively, and an interaction term for this code and hole size. An F
statistic comparing the ratio o f (the sum o f squared error attributed to the interaction o f
hole size and plant habit divided by its degree o f freedom) to (the sum o f square error for
the model divided by its degrees o f freedom) to test for differences between the Iifeformsi
Individual species height and weight responses to gap size and site for both years
were analyzed using one way anovas. Transformations were applied where appropriate.
Some species were left untransformed, some were transformed with the log
transformation, and for those which the log transformation did not normalize distributions
or equalize the variance, non- parametric Kruskall Wallace tests were used (Neter,
Wasserman, and Kutner, 1990). To make pairwise comparisons, Dunn's test, which can be
employed on unequal sample sizes and non-parametric tests Mann-Whitney tests were
used (Neter, Wasserman, and Kutner, 1990).
The light intensity data were analyzed with analysis o f variance and BonferronFs
multiple pairwise comparisons. Data were normally distributed.
Soil moisture was analyzed with Kruskall Wallace anova on ranked data, except
for the blue grama June moisture which was normally distributed. Dunn's test was used
13
for Kruskall-Wallace tests and a Bonferroni's test was used for the parametric anova.
The effect o f grasshopper control was analyzed in 1994 using the signed rank sum
test for difference in means (Neter, Wasserman, and Kutner, 1990).
SAS (1995) was used for the Box-Cox procedures and for the multi-species
ANOVAs, and Sigma Stat (1994) was used for single species ANOVA's and Kruskall
Wallace tests. Regressions and the corresponding analysis o f residual normality were run
in Minitab (1994). Soil moisture and light intensity were analyzed in SigmaStat (1994).
14
RESU LTS
Introduction
The effects o f gap size, site, year, and species on the emergence, survival, biomass,
and height o f colonizers were measured. These are presented in the first section o f the
results. Environmental conditions o f the sites, including precipitation, light intensity in the
gaps, and grasshopper herbivory, which may have contributed to the establishment
differences among gap size, site, and year, were also measured. These are presented in the
second section o f the results.
Plant Performance
Emergence Response to Gap Size,
Site. Species, and Year_________
In 1993, gap size was a significant predictor o f emergence (p=0.006. Table 2).
Emergence in Icm2 disturbances was significantly lower than in 100 cm2 and Im2 gaps
(p=0.05. Table 2). All other comparisons showed response to disturbances was similar.
In 1993, there was no difference in emergence at the two sites (p=0.649. Table 2).
Emergence within single sites could not be analyzed in 1993 because there were only tw o
responses for each site. There was no statistically significant difference between
knapweed and wheat emergence (p=0.956). The dates on which observed emergence was
greatest were 6/4/93 and 6/11/93 (Table 3). Emergence was significantly greater on these
dates than on later dates (p=O.05, Table 3), when no plants emerged.
15
In 1994, disturbance size had no significant effect on emergence (Table 2).
Emergence at the Idaho fescue site was significantly higher than at the blue grama site in
1994 (Table 2 footnote). Shepherd’s purse emergence, both total and surviving, were the
only percentages significantly different from other species (Table 4). Emergence was
higher in 1993 than 1994 (p=0.001. Table 4 footnote). Emergence differed among dates
as in 1993. That is, significantly more plants emerged between 5/26/94 and 6/4/94 than on
other dates sampled (Table 3). Survival o f cohorts from these germination dates was also
significantly higher than on other dates.
Survival Response to Gap, Site,
Species, and Year___________
In moist 1993, the Anova model relating plant survival to disturbance, species, and
site as factors was not significant (p=0.13, Table 5) Although the overall model was not
significant, site had some predictive power at (p=0.04. Table 6). Neither disturbance size
(p=0.18) nor species (p=0.89) was significant in the model.
In dry 1994, analysis o f variance on percentages o f plants which survived showed
that disturbance size, site, and species all influenced plant survival at P=O.OOOI (Table 7).
Survival was highest in the I m2 disturbances and second highest for I and 10 cm2
disturbances, and the lowest was in the 1000 cm2 gaps (Table 8). Survival was greater at
the Idaho fescue site than the blue grama (p=0.05). In 1994, seven o f the eight species
seeded survived (Table 8); no shepard's purse survived at either site.
To determine whether survival o f grasses and forbs differed, an analysis o f
16
variance treating grasses as a group and forbs as a group was used. The models predicting
survival with the factors lifeform (grass or forb), site, and gap size showed no difference
between the lifeforms (p=0.20,Table 8).
Is knapweed and wheat survival lower in dry than wet years? An Anova model
using disturbance size, site, species, and year was significant at p=0. OOOI (Table 9). Only
the Im2 disturbance had significantly higher survival than the smaller gaps. Survival was
higher in the Idaho fescue site. Knapweed and wheat survival was not significantly
different. Survival was higher in 1993 than 1994 (Table 10).
Biomass Response to Gap Size,
Site. Species, and Year_______
Two test strategies were used to determine whether biomass varied with gap size,
site, species, and year. First, weight was treated as a continuous variable, and regression
analysis was applied. Thus the log o f plant weight was regressed against site, log(gap
size) and log(gap size)2. Site was coded -I for the Idaho fescue site, and I for the blue
grama site. Years were coded as -I in 1993, and I in 1994.
Second, weight was treated as a categorical variable and analyzed with analysis o f
the variance. In finding the appropriate analysis o f variance model, all possible
combinations o f independent variables and their interactions were tried, and the best model
was chosen. Biomass was log transformed, and the factor levels were gap size, site,
species, and year.
17
1993 Biomass Regressions. In 1993, plant weight was best predicted by site, log
o f disturbance size, and log o f disturbance size squared (Table 11). For wheat and
knapweed, the R2 values explained less than 40% o f the variation.
1993 Biomass Anova. 1993 plant weight for each species was best predicted with
gap size, species, and site. Data from all plants from the 1993 field season were grouped
for one analysis o f variance (Table 12). The model was significant at p=0.0001. Gap size
was the best predictor o f plant weight (p=0.0001), and neither species (p=0.91) nor site
(p=0.26) were significant. Models with other combinations o f predictors, or which
excluded species or site were poorer. Neither species nor site affected biomass
significantly (Table 13). The I m2disturbance produced larger plants than smaller
disturbances, but response to smaller disturbance did not differ among themselves.
Data from the two sites were also analyzed separately to test for within site
differences (Tables 14 and 15). At both sites, the plants in the Im2 disturbance had
significantly higher weights than those in smaller disturbances, but responses to smaller
disturbances did not differ among themselves (Tables 16 and 17).
1994 Biomass Regressions. A summary o f the regression analysis on plant weight
from the 1994 field season for all species is presented in table 11. Plant weight increased
with gap size in a quadratic fashion for six species. Crested wheatgrass weight was
unaffected by gap size (Table 11).
Forb v. Grass Regressions. To determine whether forbs grew heavier than grasses
18
in the 1994 field season, the log o f plant weight was regressed against the log o f the
disturbance size squared, a code o f I or 0 for forb or grass respectively, and a term
reflecting the interaction o f plant habit and hole size. This interaction term was zero for
grasses, and it was the actual hole size for forbs. Other regressions were tried, and this
was best (p=0.0001. Table 18). This line explained 39% o f the variation, generally less
than the regressions describing single species, but it was intended to examine the
difference between the two lifeforms, and not to describe them singly. A partial F test
(p=0.0001) showed that forbs outperformed grasses (Table 18).
1994 Biomass Anova. Weight data from 1994 were also treated as categorical
variables to measure the effect o f gap size, that is to determine which gap sizes provide a
plants with significantly more resources. This analysis is more descriptive for comparing
specific gap sizes, sites, and species. Where both species were pooled, analysis o f
variance on the log o f plant weight against disturbance size, site, and species provided the
best model. Disturbance size and site were very good predictors o f plant weight, and
species was not (Table 19). 1,10, and 100 cm2 gaps supported equal growth (Table 20).
As gap size increased further, plant weight increased, and became significantly larger in
the I m2gap.
Sites were again examined separately to detect variation unique to each. When
pooled, weights o f all plants at the blue grama site increased gradually with disturbance
size, but no plants grew in the 1000 cm2 disturbance (Table 21). Similarly, only clover
survived in the 100 cm2 disturbance (Table 21). At the Idaho fescue site, pooled weights
in the 1000 cm2 and the I m2 disturbances were significantly greater than the smaller
19
disturbances and from one another (Tables 23 and 24).
Analysis o f variance on single species were also run on the 1994 data (Tables 1924). Analysis o f variance on single species at both sites (Table 20) indicate that weights o f
six species were equal among I, 10, or 100 cm2 disturbances. Clover was heavier in the
100 cm2 gap than smaller gaps. Clover, and wheat both grew significantly heavier in the
1000 cm2 disturbances than in the smaller disturbances. Weights did not increase further
as gap sizes increased to I m2 . The weight o f com and cheatgrass was significantly
heavier only in the I m2 gap (Table 20).
1993 v. 1994 Biomass Regressions. To compare biomass in the tw o field seasons
for wheat and knapweed, regression and analysis o f variance were again conducted.
Regression o f plant weight against site, year, log gap size, and log gap size-squared
provided the best regression line for data b n knapweed and wheat in both 1993 and 1994
(Table 11).
1993 v. 1994 Biomass Anova. Analysis o f variance using disturbance size,
species, site, and year to predict the log o f plant weight was the best model to describe the
data comparing the 1993 and 1994 field seasons (Table 25). Disturbance size was the best
predictor, and species and site were also significant predictors. Anova showed no
significant effect of year as a predictor o f biomass. Biomass from the largest disturbance
size (I m2) was significantly greater than the 1,10, and 100 cm2 disturbances, but neither
the smaller disturbances nor the largest supported significantly different growth than the
1000 cm2 disturbance (Table 26). The two sites were analyzed individually (Tables 27-
20
30). At the blue grama site plants grew significantly heavier in the Im2 disturbance (Table
28). At the Idaho fescue site disturbance sizes did not affect growth significantly (Table
30).
Height Response to Gap Size,
Site. Species, and Year_____
1993 Height Regressions. 1993 height data were analyzed using regression
analysis. A summary o f the regression analysis on plant height for all species is presented
in Table 31. In 1993, the best fit line predicted height from site, log(gap size), and
log(gap size)2. For knapweed, the regression explained 49% o f the variation in height, and
for wheat, 31% (p=0.0001 for each. Table 31).
1993 Height Anova. Analysis o f variance on 1993 height data showed disturbance
size and species to be significant determinants o f plant height (Table 32). Site was not.
Table 33 presents mean heights for each disturbance size, and the results o f pairwise
comparisons for the two species pooled and for each species individually. When wheat
and knapweed were pooled, the only disturbance size which supported increased height
was the I m2. When species were analyzed separately, wheat was significantly taller in the
Im2 gap, but no significant difference in knapweed height was seen even at this level
(Table 33). Sites were not significantly different. Wheat grew significantly taller than
knapweed (Table 33)
1994 Height Regressions. Plant height generally increased in a quadratic and
21
exponential fashion for all species except Agropyron cristatum (Table 3 1). P values were:
clover (0.0001), knapweed (0.0001), wheat (0.0001), sunflower (0.03), com (0.005),
cheatgrass (0.014) and crested wheatgrass (0.698) (Table 31).
1994 Height Anova. In 1994, plant height increased with gap size. Analysis o f
variance showed plants in I m2 gaps grew significantly larger than those in I, 10, and 100
cm2 gaps, but they were not significantly larger than those in 1000 cm2 gap's (Table 33).
Variance was also analyzed on 1994 height response to disturbance size and
species for each site separately. At the Idaho fescue site, the model, log height predicted
by gap size, site, and species, was significant, and both the disturbance size and species
factor levels were (Tables 34 and 35, but Tukey's pairwise comparisons did not show at
which factor level this occurred (Table 36 and 37). In the Bouteloua gracilis grassland,
the overall model was not significant (p=0.07, Table 34). Tukey's tests detected no
disturbance size effect (Table 36). In the Idaho fescue site, the model o f log height
predicted by gap size and species was significant at the p=0.005 level (Table 35).
Although gap size was significant in the model, pairwise tests detected no significant
differences among factor levels (Table 37).
Table 38 shows the effects o f environment on average heights in both sites, and
Tables 39 and 40 show height responses at the Idaho fescue and blue grama site
respectively. Anova tables for these analyses are presented in tables 41-43. Examination
o f these shows that the species respond differently to gap size. The forbs' height gradually
rose with increasing disturbance size, but grass height increased more slowly. This
22
contrast is seen with individual species as well (Tables 39-40). Knapweed and clover both
grew significantly taller in 1000 cm2 and Im2 disturbances, while sunflower heights were
not affected by disturbance size. None o f the grasses grew significantly taller in larger
gaps than smaller.
Grass v. Forb Regressions. Grass height increased less sharply than forb height,
and the slopes were significantly different (p=0.0001, Table 18). The best fit regression
line describing the two lifeforms height was log( height) regressed against log(gap),
log(gap)2, a code: I for forb, and -I for grass, and an interaction term for grass or forb,
and hole size.
1993 v. 1994 Height Regressions. Knapweed and wheat heights were regressed
against disturbance size, site, and year to compare their performance in wet and dry
seasons (Table 30), The regression explained 45% o f the variation for knapweed and only
23% for wheat (Table 30). Wheat plants were generally taller than knapweed plants, but
their size did not increase with disturbance size as much as knapweed (Also see Table 44).
1993 v. 1994 Height Anovas. Height increased with gap size both years. Analysis
o f variance showed that disturbance size, site, species and year all had a significant effect
on plant height (Tables 45-47). When both species were pooled, only plants in the I m 2
gap were significantly taller than those in smaller gaps. Wheat's response was the same,
but knapweed height increased more gradually, starting at the 100 cm2 gaps (Table 44).
Tukey's tests showed differences between heights with year and species, but not between
23
sites (Table 44 footnote). The sites were individually examined for differences between
1993 and 1994 using analysis o f variance (Tables 48 and 49). At the blue grama site, only
the Im2 gap grew significantly taller plants than smaller gaps (Table 48). At the Idaho
fescue site, plants in the Im2 gap grew significantly taller than those in the I, 10, and 100
cm2 gaps, but they were not significantly taller than plants in the 1000 cm2 gaps (Table
49). Height was significantly greater in 1993 at both sites.
Seed Production
Only two o f the plants studied produced seed. In 1993 neither knapweed or wheat
produced seed. In 1994, no seeds were produced at the blue grama site. At the Idaho
fescue site, one com plant produced a flower in a Im2 gap and four cheatgrass plants
produced seed. O f these four, two were grown in I cm2 disturbances, and two were
grown in 10 cm2. For com, this represents 1% o f the total amount planted, and for
cheatgrass, 6%.
Environmental Characteristics o f the Site
Precipitation
Precipitation differed between the sites and between years. In Trident, a weather
station approximately 10 km from the Bouteloua gracilis site, precipitation during 1993
exceeded normal by 19.48 cm (Table 50). In 1994, precipitation was 9.17 cm below
normal. The Montana State University weather station, which is 5 Icm from the Idaho
fescue site, recorded precipitation during 1993 19.86 cm above normal, and in 1994,
24
precipitation was 9.17 cm below normal (NOAA, 1993 and 1994). Long term
precipitation means are weighted averages calculated using data recorded from 1931 to
the present. In southwestern Montana, precipitation during 1993 was more than 200 %
greater the average for the past 30 years, and in 1994, precipitation was 50% below the 30
year average (NOAA, 1993 and 1994).
Grasshopper Herbivory
Grasshopper herbivory was severe in 1993 and less so in 1994 (Table 51).
Approximately 26% o f plants at the Idaho fescue site, and 28% o f those at the blue grama
site were at least partially eaten during the 1993 field season. In 1994, only 6% o f plants
were eaten at the Idaho fescue site, and 2% at the blue grama site. This predation was
recorded despite attempts to control it. In 1994, grasshopper numbers inside and outside
o f the study areas were compared by loop counts (Onsager and Henry, 1977). A
Wilcoxson signed rank test, the non-parametric equivalent to a paired t-test, showed that
there were significantly fewer grasshoppers in the controlled area than outside o f it (Table
52).
Light Intensity
We expected plants to grow larger and have higher survival rates in larger gaps
than smaller because disturbance releases plants from competition for light, water, and
nutrients.
The analysis o f variance was significant at p=0.0001. Pairwise comparisons
showed no significant difference in light intensity among the different disturbance sizes in
the blue grama grassland. Light levels increased with disturbance size at the Idaho fescue
25
site (Table 53). The difference results from the fact that in the Idaho fescue grassland the
vegetation was relatively tall and dense while vegetation o f the blue grama grassland was
short and open.
Soil Moisture
Soil moisture did not differ across gap sizes in June at either site, but did in July
and September (Table 54). In July at the Idaho fescue site, the IOcm2 and I m2 gaps had
significantly more moisture than the other gaps. The same month at the blue grama site,
the 1000 cm2 gaps had significantly more moisture than the other gaps. In September at
the Idaho fescue site, both the 1000 cm2 and the I m2gaps had significantly less moisture
than other gap sizes. At the blue grama site, only the 100 cm2 gap had significantly less
moisture than the other gaps. The Idaho fescue site had significantly higher soil moisture
(p=0.0001. Table 54).
26
DISCUSSION
Emergence
Peak emergence occurred in late May both years. It was concentrated in moist
1993, when 67% o f plants germinated within two weeks o f planting (Table 3). In 1994,
germination started earlier and continued for three weeks, perhaps because seeds were not
immediately exposed to enough water. Although we have no replication, these two years
represent abnormalities in Montana precipitation, with high rainfall in 1993 and low
rainfall in 1994 (Table 50). W ater may have limited emergence in 1994.
There was little effect o f gap size on emergence. In 1994, there was no overall
effect o f disturbance size on observed emergence (both grasslands and all plants pooled)
(Table 2). In 1994, when the blue grama site was considered alone, emergence was low in
the 1000 cm2 gaps (Table 2). In moist 1993 emergence was low in the I cm2 disturbance.
Emergence might have differed among gap sizes due to changes in availability o f
light, heat, or water. I) Light levels increased with increasing gap size at the Idaho fescue
site, but not at the blue grama site (Table 53). This could affect germination directly or by
heating the soil. However, at the Idaho fescue site, our results do not show the effects o f
these factors on emergence in 1994. 2) Temperature in larger disturbances is probably
warmest during the day, and cooler at night. However, no trend in emergence paralleled
this pattern. 3) We hypothesized no difference in soil water across gap sizes because soil
water is probably uniform throughout the site in moist spring. I f so, this does parallel the
emergence response.
27
Colonizing species may respond to large fluctuations in temperature in large
disturbances. This seemed to be true in some cases (Fenner, 1980; OlfE, et al, 1994), but
no so in others (Caruso, 1963). Colonizer germination was reduced when seeds were
sown into tall turf, but not short (Fenner, 1978). He thought this was an effect o f light
rather than water or nutrient preemption. Light is a trigger for seed germination and
important resource seedlings and mature plants. In a second study, Melilotus germination
in bluegrass sod was not correlated with opening size (Caruso, 1963).
Gap Size
Plant survival was not significantly affected by disturbance size in moist 1993
(Table 5). If there is a trend, raw data suggest that survival was higher in small, (I and 10
cm2) and large (I m2), than in 1000 cm2 gaps.
In 1994, survival tended to be bimodally related to gap size (Table 8). In small
gaps, (I and 10 cm2) gaps, survival is low. In larger gaps (100 and 1000 cm2) gaps it is
worse. In I m2 disturbances, survival is significantly higher. Survival in small gaps is
probably reduced by competition with established grasses. Growth is poorer still in the
1000 cm2 gaps, where exposure to heat and evaporation may be increased and competition
is still restricting growth, due to increased insolation (Table 53). Survival was best in the
Im2 gap. Here, high exposure remains, but competition is eliminated because no roots
move inward as far as the establishing seedlings.
In a mdister grassland, early seedling survival o f Melilotus in bluegrass sod was
28
correlated with opening size in openings ranging from 0.5 to 3 cm2 (Caruso, 1963). In
the two grassland we examined, gaps o f this sizes had no effect.
Simultaneous analysis o f 1993 and 1994 wheat and knapweed data indicated that
significantly higher survival percentages occurred in the Imz disturbances than in the
smaller disturbance sizes over both years (Table 10). Survival in the 100 to 1000 cm2
gaps was not significant different from that in smaller gaps.
Site
Site significantly affected survival in 1994 (Tables 8 and 9) and only slightly in
1993 (Table 5). Survival was lower at the moist Idaho fescue site than the dry blue grama
site both years. Dry conditions o f 1994 lowered survival in the Bouteloua gracilis site
more than the Idaho fescue site because o f its drier environment. Additions o f water to
blue grama grasslands benefitted plants immensely in two experiments, showing that water
is limiting in this habitat type (Weaver, 1983; Dodd and Lauenroth, 1979)
Year
Rainfall also affected survival rates. At both sites survival was lower in dry 1994,
significantly so at the blue grama site for both wheat and knapweed plants, and at the
Idaho fescue grassland for knapweed. At the Idaho fescue grassland, wheat tolerated
dryness better in 1994 than did knapweed (Table 10). These data support the fact that
water.is a limiting resource in both Idaho fescue and blue grama grasslands. Other authors
have found water to be limiting in Northern great plains grasslands as well (Weaver, 1983;
Dodd and Lauenroth, 1979).
29
Species
Wheat and knapweed survival rates did not differ in either 1993 (Table 5), or in
1994 (Table 8). However, other species' survival rates did differ in 1994. No shepherd's
purse survived. Few crested wheatgrass, com, or sunflower survived. At least 75% o f
wheat, clover, cheatgrass, and knapweed survived (Table 8).
Year and Species Interaction
Survival results were similar for these two species, although survival was higher in
1993 than 1994 and plants performed better at the Idaho fescue site. When the effect o f
seed dispersal was controlled, knapweed, a noxious weed, was no more successful than
winter wheat, a plant genetically engineered to produce grain in dryland agriculture (Table
10) .
Biomass
1993 Gap Size
In moist 1993, gaps less than Im2 provided no benefit to first year growth. This
was shown by the Anovas on gap sizes (Table 13), and more generally by regressions
using log and quadratic transformations o f the independent variables (Table 11).
Grasshopper predation may have exaggerated the benefits o f the Im2 gap, by diminishing
plant yield in 1000 cm2 gaps. Grasshopper herbivory seemed to be higher in the middle
sized 100 and 1000 cm2 gaps, than the smaller or larger gaps, although the data cannot be
statistically tested (Table 5 1). In contrast, no grasshopper damage was recorded in the
large disturbances.
30
1994 Gap Size
Log transformed biomass data from both sites in drier 1994 were pooled and were
quadraticly regressed (Table 11). We expected plant weights to increase linearly with gap
size, but yields usually did not increase dramatically until larger gaps (I m2 ). Exceptions
were wheat and clover, for which the base o f the quadratic's parabola occurred in gaps
smaller than the 1000 cm2 gap sizes.
These findings coincide with Anova results (Table 20). When all 1994 plants from
both sites were considered together in analysis o f variance, there was a gradual increase in
weight starting with the 100 cm2 disturbance, and a more precipitous increase at the Im2.
The 100 cm2 disturbance had significantly higher weights than the I cm2 gap at the blue
grama site but not at the Idaho fescue site, perhaps because biomass o f plants surrounding
the disturbances was higher at the Idaho fescue site. Neighbors around the gap probably
have reduced the effective size o f the disturbance by growing roots into it. 1000 cm2
disturbances provided significantly more resources than the 100 cm2 disturbances at both
sites in 1994 (Table 19). This was not seen in 1993 (Table 13). Perhaps in a wet year,
established plants' roots reinvaded more readily than in a dry year.
Examination of plant biomass at individual sites reveals that plant performance may
have been worse in the 1000 cm2 gaps than is evident in the anova with both sites pooled.
At the blue grama site, there was no response at the 1000 cm2 factor level (Table 21).
Because o f this, this factor level was omitted from the anova model. Had it been included,
the biomass for this gap size would have been zero grams with no variance. This is a
31
noticeable drop. Parenthetically, it should be noted that only clover survived in the 100
cm2 gap. At the Idaho fescue site, weights pooled across plants in the 1000 cm2 gaps
were significantly larger than those in the smaller gaps, but this did not occur for any
single species (Table 23). In the Idaho fescue site, plants weights began to increase at a
smaller disturbance size than at the blue grama site. Or, relating these findings to the
quadratic regression (Table 11), the base o f the parabola for gap sizes at the blue grama
site is at a larger gap size that at the Idaho fescue site.
We expect performance in gaps larger than Im2 to approximate that in the Im2.
Less resources were freed by I to 1000 cm2 disturbances than expected. This was
probably because roots o f adjacent plants quickly reinvaded the gap. This possibility is
demonstrated with root growth o f barley and wheatgrass. Five days after transplanting,
barley roots grew at a rate o f 3 cm/day. At 21 days, the rate was over 4.8 cm day
(Roddy, 1981). Both Agropyron spicatum and Agropyron desertorum roots invade
disturbances quickly (Eissenstat and Caldwell, 1989). A. desertorum roots grew to a
density o f 4 m/m2 within 20 days after disturbance, and A. spicatum grew almost 2 m/m2 in
20 days. A. desertorum is known to be more competitive than A spicatum (Eissenstat and
Caldwell, 1989).
In the competition between a colonizer and established plants, dominant species at
the site have the advantage o f well established root systems and perennating organs to
support the plant's above ground light capturing portions (Harper, 1977; Grime, 1979;
Gerry, and Wilson, 1995). Numerous studies have shown that neighbors compete in
direct response to their proximity and size. Neighborhood size affects growing season
32
performance o f target plants, but the size o f the nearest neighbor was less important
(Aguilara and Lauenroth, 1993). However, in studies examining one sided and two sided
competition on target plants, Firbank and Watkinson (1987) found that crowding, that is
small neighborhood size, accounted for only a small amount o f variation in plant size.
When crowding was considered with emergence time, the variance accounted for in
multiple regression was increased by 50%. Whether the site was homogenous or
heterogenous outweighed the importance o f competition in the establishment of
developing seedlings (Fowler, 1988). The presence o f nearby adult neighbors inhibited
seedling growth in some cases and not in others (Fowler, 1988).
1993 Site
In 1993, when water was not limiting, production was similar In the blue grama
aridisol and the Idaho fescue mollisol, which is normally more productive. In 1994, when
water was limiting, growth was better at the Idaho fescue site than the blue grama. 1994
was slightly drier than average for the region. Logan normally gets less precipitation than
Bozeman, and soil moisture was lower at the blue grama site (Table 54).
1994 Site
In 1994, production was greater at the Idaho fescue site (Table 20). In 1994,
emergence, survival, biomass, and height responses all indicate this, contrary to 1993
results, when sites were equal. As expected plants were larger and had higher survival
percentages at the w etter Idaho fescue site. Exceptions occurred in the I and 10 cm2
gaps, due to strong root competition. At the blue grama site, plant sizes did not increase
33
with increasing disturbance size. At the Idaho fescue site, plant size did increase with
disturbance size.
Grass v. Forb
In 1994, forb weights increased more rapidly with increasing disturbance size than
did grass weights (Table 18). We hypothesize that forb success is likely because root
competition between diffuse rooted grasses was more intense than between grasses and
taprooted forbs. In the I cm2 disturbances, plant weights were similar (Table 20). There
is no competitive advantage to either lifeform here. Root architecture which invades areas
where other lifeforms do not predominate access water or nutrients depleted in areas
where roots systems compete (Lynch, 1995). This is consistent with the observation that
most noxious range weeds are forbs.
Seed Size
Seed size had no affect on final plant biomass. For example, while none o f the
lightest seed, shepherd's purse established, the second smallest, clover, had one o f the
highest biomasses and survival rates o f all species at both sites (Table 19). Sunflower and
corn, the heaviest seeds, grew very little in the blue grama site, and only slightly at the
Idaho fescue site. This agrees with previous literature. For example, seed size did not
affect colonizer success in a study o f ruderal invasion o f gaps in short and tall tu rf grass
(Fenner, 1978).
34
Year
When knapweed and wheat growth were compared over a two year period,
performance differed between grassland types in the dry year but not in the wet. Biomass
was significantly higher in the Idaho fescue site in the second year (Table 20). This shows
that water was limited in the dry year. When water was adequate, colonizer production
was the same for both sites. However, in dry conditions, production was higher in the
moister Idaho fescue site.
Height
Height responses to gap size did not increase significantly until the largest gap
sizes (Tables 33 and 38). It is presumably determined by two forces: light and
photosynthesis. Based on light, plants are expected to etiolate as the stand closes. Based
on photosynthesis, height will decline as density increases; due to reduced resources.
Thus, height measurements are not a simple measure o f success. They are also
confounded by the influence that competition may have on the physiology o f the plant. In
smaller disturbances, it is beneficial for a plant to expend its energy growing tall quickly.
Gap
In 1993, only the Im2 disturbance provided significant benefit for plant height
when both sites were pooled (Table 33). Smaller disturbances behaved similarly to one
another. When considered alone, knapweed did not grow significantly larger in the Im2
gap. Small n (only 2 responses) for the I m2 disturbance may explain why the pairwise
35
comparisons failed to detect a difference for knapweed in the Im2 disturbance (Table 33).
In 1994, plants in the 1000 cm2 disturbance were significantly taller than in the
smaller disturbances (Table 38). Knapweed and clover showed significant height increases
in the 1000 cm2 gaps, and this may have swayed the statistics when all plants were pooled.
Site and Year
In 1993, plant height was unaffected by site, while in 1994, it was affected (Tables
33 and 38). Due to high rainfall in 1993, water was plentiful in both places, thus there
were no differences in height. In 1994, there was little rainfall, and the lack o f rainfall was
even more pronounced at the blue grama site than at the Idaho fescue.
Seed Production
Seed production is a standard measure o f a plant's fitness. Three o f the plants
studied—winter wheat, knapweed, and yellow sweet clover were biennial or perennial,
crested wheatgrass—do not produce seed during, the first growing season and cannot be
evaluated by this criterion. Sunflower, cheatgrass, com, and shepard's purse were all
capable o f seed production during the first growing season, but no sunflower, only one
com plant, and a few cheatgrass plants produced seed, making it the most successful o f
the annuals.
36
CO NCLUSIO N
Disturbance size plays an important role in colonizer success. Plants were more
successful in larger gaps than smaller. I) We hypothesized that germination would not
vary with gap size because o f abundant water in the spring. In dry 1994, disturbance size
had no effect on emergence. In wet 1993, however, the smallest disturbance (I cm2) had
significantly lower emergence than larger disturbance sizes. 2) Gap size had no effect on
survival in 1993, but did in 1994, when survival was lowest in the 1000 cm2 gaps. 3)
Biomass increased with gap size in both years. The I m2 gap supported more growth than
the smaller gaps in 1993; The latter did not differ among themselves. In 1994, the I m2
gaps were significantly larger than all other gaps. Plants in the 1000 cm2 gaps were
smaller than these, but significantly larger than plants in the I to 100 cm2 gaps. 4) A
monotonic increase was seen in plant heights in 1993, but in 1994, plant size only became
significantly greater in the Im2 gaps.
Forbs produced more biomass than grasses and especially so in the 1000 cm2 to I
m2 disturbances. Both behaved similarly in the largest I m2 gaps. W e suggest that
taprooted forbs avoid competition with sod to a greater degree than diffuse rooted
grasses.
Plant response to increasing gap size might be due to reduction in competition, for
materials such as water and nutrients, or physical factors, (light, heat, or wind). Attempts
to compare water availability and temperatures across disturbance sizes were largely
unsuccessful. Light intensity increased with gap size in the Idaho fescue grassland, and
37
not in the blue grama. The denser canopy at the Idaho fescue site apparently shaded the
smaller disturbances. But in the blue grama site, gaps in the canopy were so large that
difference in light penetration to the soil surface could not be measured.
Plant performance at the blue grama site was expected to be poorer than at the
Idaho fescue site since water and nutrients are lower at the blue grama site. Emergence
was unaffected by site in wet 1993, but in dry 1994, it was higher in the Idaho fescue site.
As expected, survival was significantly higher in the Idaho fescue site both years. Biomass
was not significantly affect by grassland type in 1993, when water needs were satisfied,
but in 1994, when water was lacking, plants were heavier at the Idaho fescue site. The
same was true for height. In 1993, height was unaffected by grassland type, but in 1994,
plants grew taller in the Idaho fescue site.
By all measures: emergence, survival, biomass, and height, both wheat and
Icnapweed performance was significantly higher in moist 1993 than dry 1994.
Species behaved differently in 1994 but not in 1993. I) Germination differed little
among species. Survival differed more. Wheat and knapweed emergence and survival did
not differ in either 1993 or 1994. In 1994, all species had similar rates o f emergence,
except shepard's purse which had significantly lower emergence rates than wheat. 2)
Among all plants in 1994, clover had the highest survival rates, wheat, knapweed, com,
and cheatgrass were intermediate, sunflower had low survival percentages, and crested
wheatgrass had the lowest. Wheat, knapweed, com, and cheatgrass were not significantly
different from one another. Survival did not differ between wheat and knapweed.
LITER A TU R E CITED
39
LITER A TU R E CITED
Aguilera, M O., and W.K. Laurenroth. 1993. Neighborhood interactions in a natural
population o f the perennial bunchgrass Bouteloua gracilis. Oecologia 94: 595602 .
Aguilera, M O., and W.K. Laurenroth. 1993. Seedling establishment in adult
neighborhoods- intraspecific contraints in the regeneration o f the bunchgrass
Bouteloua gracilis. Journal o f Ecology 8 1: 253-261.
Amo, S E. and G E. Gfuell. 1983. Fire History o f the Forest-grassland ecotone in
southwestern Montana. Journal o f Range Management. 36: 322-336.
Campbell, B.D., J.D. Grime, and J.M.L. Mackey. 1991. A trade-off between scale and
precision in resource foraging. Oecologia 87: 532-538.
Caruso, J.L. 1963. Early Seedling Survival o f Melilotus in bluegrass sod. Ecology,
51(3): 553-554.
Coffin, D.P., and W.K. Laurenroth. 1988. The effects o f disturbance size and frequency
on a shrotgrass plant community. Ecology 69(5): 1610-1617.
Coffin, D.P., and W.K. Laurenroth. 1991. Effects o f competition on spatial distribution
o f roots o f blue grama. Journal o f Range Management 44(1): 68-71.
Coffin, D.P., and W.K. Laurenroth. 1990. A gap dynamics simulation model of
succession in a semiarid grassland. Ecolgical modelling 49: 229-266.
Coffin D.P., and W.K. Laurenroth. 1989. Small scale disturbances and successional
dynamics in a shortgrass plant community : interactions o f disturbance
characteristics. Phytologia 67(3): 258-286.
Coffin D.P., and W.K. Laurenroth, 1994. Successional dynamics o f a semiarid
grassland: effects o f soil texture and disturbance size. Vegetation 110: 67-82.
Daubenmire5R. 1978. Plant geography. Academic Press, New York.
Dodd, I , and W. Lauenroth, 1979. Analysis o f the response o f a grassland ecosystem to
stress. In French, N. Perspectives in grassland ecology. Springer, NY. 204
pages.
40
Eisenstat, D.M., and M.M. Caldwell. 1989. Invasive root growth into disturbed soil o f
two tussock grasses that differ in competitive effectiveness. Functional Ecology,
3: 345-353.
Fenner, M. 1978. A comparison o f the ability o f colonizers and closed tu rf species to
establish from seed in artificial swards. Journal o f Ecology 66:953-963.
Fenner, M. 1980. Seed characteristics in relation to succession. In: Colonization.
succession, and stability: the 26th symposium o f the British Ecological Society
held jointly with the Linnean Society o f London. Blackwell Scientific, Boston,
MA, 103-114.
Firbank, L.G., and A R . Watkinson. 1987. On the analysis o f competition at the level o f
the individual plant. Oecologia 71:308-317.
Fowler, N. 1988. What is a safe site?: Neighborhood, litter, germination date and patch
effect. Ecology 69(4):947-961.
Friend, D.T.C. 1961. A simple method o f measuring integrated light values in the Field.
Ecology. 42(3): 577-580.
Geriy, A.K., and S.D. Wilson. 1995. The influence o f initial size on the competitive
responses of six plant species. Ecology 76(1): 272-279.
Gould, F.W., and R.S. Shaw. 1983. Grass Systematics. Texas A&M University Press,
Austin TX. 397 pp..
Grime, J.P. 1979. Plant Strategies and Vegetation Processes. John Wiley and Sons, New
York. 222 pp.
Hitchcock, C.L., and A. Cronquist. 1973. Flora o f the Pacific Northwest. University o f
Washington Press, Seattle, W A 730 pp.
Hook, P B., TC. Burke, and W.K. Laurenroth. 1991. Heterogeneity o f soil and plant N
and C associated with individual plants and openings in N orth American shortgrass
steppe. Plant and Soil 138: 247-256.
Hook, P.B., and W.K. Laurenroth. 1994. Root system response o f a perennial
bunchgrass to neighborhood-scale soil water heterogeneity. Functional Ecology
8: 738-745.
41
Hoojc, P B., and W.K. Laurenroth. 1994. Spatial patterns o f roots in a semiarid
grassland: abundance o f canopy openings and regeneration gaps. Journal o f
Ecology 82: 485-494.
Lieberman5M. and D. Lieberman. 1989. Forests are not just swiss cheese: Canopy
stereogeometry o f non-gaps in tropical forests. Ecology 70(3): 550-552.
Lynch, J. 1995. Root architecture and plant productivity. Plant Phsysiology 109: 7-13.
McConnaughay, K.D. and F.A. Bazzaz. 1987. The relationship between gap size and
performance o f several colonizing annuals. Ecology 68(2): 411-416.
McConnaughay, K.D. and F.A. Bazzaz. 1990. Interactions among colonizing annuals: Is
there an effect o f gap size? Ecology 71(5): 1941-1951.
Minitab Inc. 1995. Minitab for Windows. State College, PA.
National Oceanic and Atmospheric Administration. 1992. Climatological Data:
Montana. Department o f Commerce, Ashville, NC. 95:13.
National Oceanic and Atmospheric Administration. 1993. Climatological Data:
Montana. Department o f Commerce, Ashville, N C. 96:13.
National Oceanic and Atmospheric Administration. 1994. Climatological Data:
Montana. Department o f Commerce, Ashville, NC. 97:13.
Neter, J., W. Wasserman, and M. H. Kutner. 1990. Applied Linear Statistical Models:
Third Edition. Irwin, Boston, MA. 1181pp.
OlfF, H., D M. Pegtal, J.M. vanGroendendael, and J.P. Bakker. 1994. Germination
strategies during grassland succession. Journal o f Ecology 83: 69-77.
Onsager, LA., and J.E. Henry. 1977. A method for estimating the density o f rangeland
grasshoppers in experimental plots. Acrida 6: 231-237.
Pickett, S.T.A., and P S . White. 1985. The ecology o f natural disturbance and patch
dynamics. Academic Press, Inc. Orlando, FE. 472 pp.
Platt, W.J. 1975. The colonization and formation o f EquiHbrium plant species
associations on badger disturbances in a tail-grass prarie. Ecological Monographs
45: 285-305.
42
Reader, R J ., S.D. Wilson, J.W. Belcher, P A . Kennedy, D. Tilman, E.C. Morris, J.B.
Grace, H. Olf, R. Turkington, E. Klein, Y. Leung, B. Shipley, R. van Hulst, M.E.
Johansson, C. Nilsson, J. Gurevitch, K. Grigulis, and B E. Beisner. Plant
competition in relation to neighbor biomass: and intercontinental study with Pod
pratensis. Ecology 75(4): 1753-1760.
Reice, S R . 1994. Nonequilibrium determinants o f biological community structure.
American Scientist 82: 424-435.
Sala, G E., W.K. Laurenroth, and W.J. Parton. 1992. Long-term soil water dynamics in
the shortgrass steppe. Ecology 73(4): 1175-1181.
SAS. 1995. The SAS system for windows. SAS Institute, Inc. Cary, NC.
SigmaStat. 1994. Statistical computer software. Jandel Scientific.
Tilman, D. 1982. Resource Competition and Community Structure-. Princeton University
Press, Princeton, NJ. 296 pp.
Tilman, D. 1988. Plant strategies and the dynamics and structure o f plant communities.
Princeton University Press Princeton, NJ. 360 pp.
Umbanower, C E. 1991. Early Pattern o f revegetation o f artificial earthen mounds in a
northern mixed prairie. Canadian Journal o f Botany 70: 145-150.
Weaver, T. 1979. Changes in soils along a vegetational (altitudinal) Gradient o f the
Northern Roclcy Mountains, pp. 65-280. In: C Youngberg ed. Proceedings o f the
Fifth N orth American Forest Soils Conference. Soil Science Society o f America.
Madison WI. 172 pages.
Weaver, T. The yeild response o f high plains grasslands to water enrichment, three
phases, pp. 77-94. In Holman, L. and G. Knudson. The high plains cooperative
program: 1981-1983. Montana Department o f Natural Resources. Helena, MT.
172 pages.
Winer, B I. 1962. Statistical Principals in Experimental Design. McGraw-Hill, New
York, NY. 92-96.
APPENDIX
44
Table I. Characteristics o f plants treated.
scientific name
common name
abbrev.
seed
weight
rooting
pattern
regen.
strategy
year
Agropyron
cristatum
crested
wheatgrass
agcr
0.0020
diffuse
perennial
94
Bromus tectorum
cheatgrass
brte
0.0020
diffuse
annual
94
Triticum aestivum
winter wheat
trae
0.0200
diffuse
biennial
93
94
Zea mays
corn
zema
0.1600
diffuse
annual
94
Capsella bursapastoris
shepard's purse
cabu
0.0001
tap
annual
94
Melilotus
officinalis
yellow sweet
clover
meof
0.0020
tap
biennial
94
Centaurea
maculosa
knapweed
cema
0.0020
tap
biennial
93
94
Helianthus
annuus
sunflower
bean
0.1400
tap
annual
94
grasses
forbs
45
Table 2. Emergence (%) o f all species at five gap sizes.
I cm
IO cm
IOO cm
1000 cm
Im
25
15
5, IO2
5, IO2
L42
94 both sites3
74 a4
81a
54 a
43 a
59 a
93 both sites
59a
69 ac
80 c
72 ac
80 c
94 Bogr56
15 a
16 a
9.5 ah
4b
8 ab
93 Bogr
80
84
100
90
100
94 Feid
76 a
81a
64 a
65 a
78a
93 Feid
66
90
100
90
100
#
possible3
’Possible responses, percentages based on this number.
-In 1993. the first number was possible, in 1994, the second.
3Analyzed with Kruskall-Wallace Anova on ranked data (1993 both sites, p=0.008; 1994 both sites,
p=0 05; blue grama 1994. p=0.006; Idaho fescue. 1994. p=0 479).
’Letters indicate Dunn's pairwise comparison results (p=0.05). These are intended to be compared
across rows, and not down columns.
5Sites were significantly different in 1994, not in 1993 (p=0.005).
6 Bogr= Blue grama, Feid= Idaho fescue
1Due to insufficient sample sizes, ANOVAs could not be run on individual site's gap sizes in 1993.
Mann-Whitney tests showed that overall, sites were not significantly different in 1993 (p=1.000).
46
Table 3. Emergence (%) and survival (%) at each date and site.
d a te
4 /2 3 -
5 /2 6 -
6 /1 1 -
7 /1 -
7 /1 5 -
o b s.
4 /3 0
6 /4
6 /1 8
7 /9
7 /2 6
94
tot.
H%3
S% '
L0Zo
S 0Zo
E0Zo
S%
E 0Zo
S%
E%
S%
E0Zo
S%
5 a3
3 a
35
14 b
23 b
9bo
4 a
3a
0 .1 a
Oa
67
29
9b
12 a b
0 .6 5
Oa
0 a
0 a
0 a
59
11
b o th
b
s ite s
94
3 a
I a
44 b
B o g r"5
94
ab
7 ab
5ab
25 b
19b
34 b
17b
8 ab
5ab
Ob
Ob
74
46
84 a
73 a
IO a
IO a
Ob
Ob
Ob
Ob
94
83
91
74
3
3
0
0
0
0
91
77
78
72
16
16
0
0
0
0
94
87
F e id
93
b o th
s ite s
93
B o g r fi
93
F e id
1Percent of plants that emerged.
-Percentage of plants that surv ived.
3Analyzed with Kniskall-Wallace Anova on ranked data (1993 both sites, p=0 0043; 1994 both
sites, p=0.0001; blue grama 1994, p=0.0001; Idaho fescue, 1994, p=0 0013).
"Letters indicate Dunn’s pairwise comparison results (p=0.05). Compare across rows, and not down
columns. Do not compare total percent column with surviving percent column.
5Bogr= Blue grama, Feid= Idaho fescue
^Insufficient data to run ANVOAs in 1993 single site data.
4
47
Table 4. Total emergence (%) for each plant.
cabu'
meof
cema
hean
agcr
bite
trae
zema
N2
64
64
5 1 .6 4 3
64
64
64
5 1, 64:
64
94 both
sites4'5
22 a6
86 b
58 ab5
58 ab
54 ab
62 ab
95 b
64 b
93 both sites
94 Bogr
83 a
35 a
82 ab
93 BogC 9
94 Feid
93 Fcid
38 ab
84 a
31 b
51 ab
45 ab
76
9a
91 b
77 b
90
93 b
40 ab
86
85 b
57 ab
78 b
97 b
89 b
82
lCabu=Shepard1S purse, meof=colvcr. ccma=knapweed, hcan=sunflovver. agcr=crestcd wheatgrass,
brte=cheatgrass, lrae=wheat, zema=com.
'Possible responses, percentages based on this number.
Mn 1993, the first number was possible, in 1994, the second.
4Analyzed with Kruskall-Wallace Anova on ranked data (1993 both sites. p=0.9698; 1994 both
sites, p=0.OOOI; blue grama 1994, p=0.0001; Idaho fescue, 1994, p=0.0001).
5Wheat emergence was not significantly different in 1993 and 1994 (p>0.05). knapweed emergence
was significantly different in the two years (p=0.05).
6Letters indicate Dunn's pairwise comparison results (p=0.05). Compare across columns, and not
down rows.
Sites were significantly different in 1994, not in 1993 (p=0.05).
^Insufficient data from individual sites in 1993 to analyze variance for single species.
9Whcat had significantly higher emergence in 1993 (p=0.05)
48
Table 5. Effect o f gap size, site and species on plant survival (%), 1993
Icm
IOcm
100
cm
1000
cm
Im
25'
Avg3
15
Avg
5
Avg
5
Avg
I
Avg
tot6
Cema3B45
F
64
92
78
87 67
77
60
100
80
60
100
80
100
100
100
a
trae B
F
92
88
90
93 87
90
60
100
80
40
80
60
100
100
100
a
total B7
F
78
90
84 a
90
77
84 a
60
100
80 a
50
90
70 a
100
100
100a
3Average o f the two sites.
3Cema^knapweed. trae=wheat
4 B= blue grama, F= Idaho fescue
5Sites significantly different at p=0.05
6Total surv ival for each species compared. Tukey's test (p=0.05).
Anova (p=0.13) and Tukeys pairwise comparisons (p=0.05).
Table 6. Analysis o f variance on wheat and knapweed survival data, 1993 ..
Source
DF
Sum of Squares
Mean Square
F Value
Pr>F
Model
6
3121
520
2.06
0.13
Error
13
3277
252
Corrected
Total
19
6399
Source
DF
Type III SS
Mean Square
F Value
Pr>F
Hole
4
1868
467
1.85
0.18
Spp
I
5
5
0.02
089
Site
I
1248
1248
4.95
0.04
R Square
CV
Root MSE
Mean
0.49
19.02
15.88
83.50
49
Table 7. Analysis o f variance on survival data from 1994
Source
DF
Sum o f Squares
Mean Square
F Value
Pr^F
Model
11
57183.56
5198.51
15.08
V
Error
58
19994.51
344.73
Corrected
Total
69
77178.07
Source
DF
Type 111 SS
Mean Square
F Value
Pr>F
Hole
4
13042.286
3260.57
9.16
0.0001
Spp
6
17523.77
2920.63
8.47
0.0001
Site
I
26617.50
26617.50
77.21
0.0001
R Square
C V.
Root MSE
Mean
0.74
52.09
18.57
35.64
'01
50
Table 8. Effect o f gap size, site, and species on plant survival (%), 1994
Ic
10
100
1000
1000c
tot
25'
Avg123
15
Avg
10
Avg
10
Avg
10
Avg
m cof B5
F
52
88:
70
53
60
57
20
90
55
20
90
45
75
100
88
a6
trae
B
F
16
76
46
7
73
40
0
90
45
0
50
25
75
100
88
ab
ccma B
F
16
60
38
27
73
50
0
60
30
0
10
75
75
75
abc
zema B
F
0
64
32
13
93
53
0
50
25
0
60
30
0
75
38
acd
brte
B
F
0
44
22
7
33
20
0
40
20
0
30
15
50
100
75
acd
hean B
F
0
8
4
8
53
31
0
20
10
0
0
0
25
50
38
cd
agcr B
F
8
52
30
13
33
23
0
20
10
0
0
0
25
0
13
d
Ibrb4 B
F
23
52
37
18
62
46
7
57
32
0
37
18
58
75
67
e
grass B
F
6
59
33
10
58
34
0
50
25
0
35
18
38
69
53
e
total B6
F
12
49
35 ab
16
52
39 a
3
46
28 ab
0
31
18b
41
63
59 c
20
1 Number of responses possible, percentages based on this number.
2 B= blue grama, F= Idaho fescue
3Average o f the two sites.
4Cabu=shepard's purse. mcof=colver, ccma=knapvveed, hean=sunflower, agcr=crested wheatgrass,
brte=cheatgrass, trae=wheat, zema=com.
"Grasses and forbs are not included in the total average, and should not be compared to individual
species percentages in the last cloumn.
6Anova (p=0.0001) and Tukcvs pairwise comparisons (p=0.05).
51
Table 9. Analysis o f variance on survival data, 1993 and 1994.
Source
DF
Sum of Squares
Mean Square
F Value
Pr-F
Model
7
3409.68
4865.67
15.03
0.0001
Error
32
10359.10
323.72
Corrected
Total
39
44418.78
Source
DF
Type III SS
Mean Square
F Value
Pr>F
Hole
4
9180.40
2295.10
7.09
0.003
Spp
I
207.03
207.03
0.64
0.43
Site
I
9597.03
9579.03
29.59
0.0001
Year
I
15093 23
15093.23
46.62
0.0001
R Square
CV
Root MSE
Survival Mean
0.77
28.08
17 99
64.08
Table 10. Effect o f site and year on survival (%) o f knapweed and wheat, 1993 and 1994
Icm
10
cm
100
cm
1000
cm
Im
5,10'
5,10'
1,4'
possible
25'
Axg2
15'
cema393 B
F4
64
92
78
87
67
77
60
100.
80
60
100
80
100
100
100
a
cema 94 B
F
16
60
38
27
73
50
0
60
30
0
20
10
75
75
75
b
trae 93
B
F
92
88
90
93
87
90
60
100
80
40
80
60
100
100
100
a
trae 94
B
F
16
76
46
7
73
40
0
90
45
0
50
25
75
100
88
b
44b
80
90
" i
#
17
59b
53
64 b 20
47
63 b
B
53
83
75
79
F
'Number of responses possible, percentages based on this number
3Ax erage of the two sites.
3Cema=knapweed. trac=whcat
4 B= blue grama. F= Idaho fescue
5Sites significantly different at p=0.05
6Total survival for each species compared
7Anova (p=0.0001) and Tukeys pairwise comparisons (p=0.05).
total
52
Table 11: Relationship o f plant weight (g) to site, year, and gap size.'
trans
form2
Y
intercept
site
coef.
meoP 94
2
-1 9 6
hean
94
2
bite
94
year
coef.
log
(hole)
coef.
log
(hole)2
coef.
R2
p value
0.07
0 27
0.07
0 62
0.0001
-1.40
0.01
0 36
0.04
0.79
0.0001
2
-2 82
-0.27
0 26
0.05
0.73
0.0001
agcr 94
2
-2.44
0.18
0.13
-0.07
0.09
0.0001
zema 94
2
-1.24
0.10
-0.1
0.10
0.48
0.0001
trae 93
2
-2.03
-0.13
0.04
0.11
0 38
0.0001
trae 94
2
-1.58
0.02
-0.49
0 22
0.63
0.0001
trae 93 &
94
2
0.15
0.01
-0.32
0.16
0.57
0.0001
cema 93
2
-1.78
-0.01
-0.34
0.21
0.39
0.0001
cema 94
2
-2.10
0.15
-0.04
0.11
0.51
0.0001
cema 93
& 94
2
-1.96
0.03
-0.20
0.16
0.44
0.0001
-0.02
-0.20
1All possible regression equations were tried, these fit the data the best.
2I= no transformation. 2= log 10 transformation.
3Cabu=shcpard's purse, mcof=colver, ccma=knapweed, hean=sunflower, agcr=crcsted wheatgrass,
brte=cheatgrass, trac=wheat, zcma=com.
53
Table 12. Analysis o f variance on wheat and knapweed weight from both sites in 1993.
Source
DF
Sum o f Squares
Mean Square
F Value
Model
6
16.560
2.760
17.81
Error
13
2.015
0.155
Corrected
Total
19
18.574
Source
DF
Type III SS
Mean Square
F Value
Pr>F
Hole
4
16.344
4.086
26 37
0.0001
Spp
I
0.002
0.002
0.01
0.9100
Site
I
0.213
0.213
1.3800
0.2620
R Square
CV
Root MSE
Log (wt) Mean
0.891
-36.287
0.394
-1.085
Jj
Table 13. Plant weights (g) at both sites, 1993'.
100cm
10cm
Icm
Im
I OOOcrn
cerna2-3,4
0.02
a5
0.02
a
0.07
a
0.30
a
3.21
a
trae
0.02
a
0.03
ab
0.03
ab
0.05
ab
8 31
b
total6
0.02
a
0.03
a
0.04
a
0.19
a .
5.76
b
'Sites are not significantly different (p=0.05).
:cema=knapweed, trae=wheat
2Species are not significantly different (p=0.05).
4Single species analyzed with Kruskall-Wallace analy sis o f variance on ranks (wheat. p=0 0029;
knapweed, p=0.0010t and Dunn's pairwise comparisons (p=0.05).
'Letters indicate pairwise comparison results (p=0.05). Compare across rows, not down -uiumns.
6Anova (p=0.0001) and Tukev's pairwise comparisons (p=0.05).
54
Table 14. Analysis o f variance on wheat and knapweed weight, blue grama site in 1993.
Source
DF
Sum o f Squares
Mean Square
F Value
Pr>F
Model
5
41.5700
8.310
14.460
0.010
Error
4
2.300
0.57
Corrected
Total
9
43.870
Source
DF
Txpc III SS
Mean Square
F Value
Pr>F
Hole
4
40.960
10.240
17.810
0.008
Spp
I
0.610
1.060
0.360
‘ 0.61
R Square
CV
Root MSE
Log (xvt) Mean
0.95
-27.72
0.76
-2.74
Table 15 . Analysis o f variance, wheat and knapweed weights, Idaho fescue site, 1993.
Source
DF
Sum o f Squares
Mean Square
F Value
Pr>F
Model
5
51.900
10.38
26.310
0.004
Error
4
1.580
0.390
Corrected
Total
9
53.480
Source
DF
Txpe 111 SS
Mean Square
F Value
Pr>F
Hole
4
51.040
12.760
32.35
0.003
Spp
I
0.86
0.8600
2.1800
0.210
R Square
CV
Root MSE
Log (xxt) Mean
0.97
-27.79
0.63
-2.26
55
Table 16. Wheat and knapweed weights (g) at the blue grama site, 1993.
Icm
IOcm
100cm
IOOOcrn
10000cm
cema1'3
0.03
a4
0.02
a
0.01
a
0.03
a
1.73
a
trae
0.02
a
0.04
a
0.05
a
0.01
a
7.86
a
total5
0.02
a
0.04
a
0.05
a
0.37
a
3.19
b
1Species are not significantly different (p=0.05).
:ccma=knap\vecd. Irae=Whcat
1Letters indicate pairwise comparison results (p=0.05). Compare across rows, not down columns.
4Single species analyzed with Kruskall-Wallace analysis of variance on ranks (wheat, p=0.0011;
knapweed. (p=0.0613) and Dunn's pairwise comparisons.
5Anova (p=0.010) and Tukey's pairwise comparisons (p=0.05).
Table 17. W heatanc knapweed weight (g) at the Idaho 'escue site, 1993.
I0000cm
IOOOcrn
100cm
10cm
Icm
cema1,3'3
0.01
a4
0.03
a
0.07
a
0.45
a
4.68
a
trae
0.01
a
0.02
a
0.02
a
0.08
a
8.75
a
0.74
b
0.37
a
0.21
a
0.04
a
0.02
a
total5
'Species are not significantly different (p=0.05).
:Cema=knapweed. trac=whcat
’Letters indicate pairwise comparison results (p=0.()5). Compare across rows, not down columns.
’Single species analyzed with Kruskall-Wallace analysis of variance on ranks wheat p=0.0014and
Dunn's pairwise comparisons, pooled species analyzed with anova and Tukey 's pairwise
comparisons.
5Anova (wheat, p=0.014; knapweed, p=0.0004) and Tukey's pairwise comparisons (p=0.05)
56
Table 18. Grass and forb height and weight in relationship to gap size.'
tf-
Y in t
Ig(hole)2
coef.
life form
coef.3
lifeform/
hole int.
R2
p value
F*
Wt
2
-1.900
0.082
-0.026
0.042
0.403
0.0001
7.5
ht
2
0.743
0.014
0.129
0.019
0.170
0.0001
8.03
i n value
,01
I 0.0001
1Numerous regression equations were tried, this fits the data the best.
2I= no transformation. 2= log 10 transformation.11= no transformation. 2= log 10 transformation.
3 X is zero for grasses, and actual gap size for (orbs
4Partial F test for slope difference (Netcr. Wasscrman, and Kutner, 1993).
57
Table 19. Analysis o f variance, plant weights (log transformed), both sites, 1994.
Source
DF
Sum of Squares
Mean Square
F Value
Pr>F
Model
11
3.816
3.347
18.28
0.0001
Error
38
6.957
0.183
Corrected
Total
49
43.772
Source
DF
Tvpc III SS
Mean Square
F Value
Pr>F
Hole
4
20.145
5.036
27.51
0.0001
Site
6
13.740
2.290
12.51
0.0001
Spp
I
0.357
0.357
1.95
0.171
R Square
CV
Root MSE
Log (wt) Mean
0.841
-35.88
0.428
-1.193
Table 20. Effect o f gap size on plant weights (g), both sites, 1994.'
100cm
10cm
Icm
10000cm
IOOOcrn
meof2
0.018
a
0.040
a
0.122
ab
0.355
b
2.654
b
cema
0.010
a3'4
0.010
a
0.017
a
0.360
a
0.514
a
bean
0.087
a
0.150
a
0.229
a
6.261
a
agcr
0.004
a
0.000
a
0.008
a
0.001
a
bite
0.006
a
0.010
a
0.013
a
0.050
a
0.290
b
trac
0.029
a
0.020
a
0.020
a
0.171
b
2.740
b
zema
0.050
a
0.060
a
0.073
a
0.190
a
13.865
b
forbs5
0.020
0.040
0.100
0.360
2.510
grass
0.030
0.030
0.030
0.160
1.400
total5
0.020
a
0.040
a
0.070
ab
0.251
be
2.000
'Sites were significantly different at p=0.05.
2Meof=Clover, cema=knapwced. hcan=sunflower, agcr=crcsted wheatgrass, brte=cheatgrass.
trae=wheat.zema=com.
■Letters indicate pairwise comparisons results (p=0.05). Compare across rows, not columns.
lKruskall -Wallace analysis of variance on ranks (Clover, p=0.0()01; knapweed. p=0.0092;
sunflower. p=0.0498; crested wheatgreass. p=0.3278; cheatgrass. p=0.0012; wheat. p=0.0009;
com, p=0.0073). Dunn's pairwise comparisons used for all single species comparisons.
sAnova (p=0.0001). Tukev s pairwise comparisons (p=0.05). Weights were log transformed.
58
Table 21. Effect o f gap size on plant weights (g), blue grama site, 1994.
Icm
IOcm
IOOcm
cema1
0.012
a2
0.013
a
meof
0.021
a
0.036
a
hel3
trae
0.212
IOOOcrn
10000cm
a
0.169
0.028
a
0.013
zema
0.072
bite
0.003
agcr
0.011
0.004
forbs
0.020
0.050
grass
0.030
0.030
total4
0.021
a
0.042
0.853
b
1.518
b
1.801
a
0.709
a
0.090
0.220
0.260
0.320
ab
0.212
be
0.736
C
1Meof=clovcr: cema=knapwced. hcan=sunflovvcr, agcr=crested whcatgrass. brte=chcatgrass,
trae=\vheat=zema=com.
2Single species weights analyzed w ith one w ay analysis o f variance (clover. p=0.0001; knapweed,
p=0.0005; wheat, p=0.036) and Student-Neuman-Kuels pairwise comparisons. All weights log
transformed.
’Crested whcatgrass, com. sunflower, and cheatgrass had too few responses to analyze.
'Total species analysis used log transformed weights and analysis of variance (p=0.06) and Tukey's
painvise comparisons.
Table 22: Analysis o f variance on plant weights (log transformed), blue grama site, 1994.
Source
DF
Sum of Squares
Mean Square
F Value
Pr>F
Model
9
63.68
7.08
3.16
0.06
Error
8
17.89
2.24
Corrected
Total
17
81.56
Source
DF
Type 111 SS
Mean Square
F Value
Pr>F
Hole
3
27.12
9.04
4.04
0.05
Spp
6
35.24
5.87
2.63
0.10
R Square
CV
Root MSE
Log (wt) Mean
0.78
-48.92
1.50
-3.06
59
Table 23. Effect o f gap size on plant weights (g) at the Idaho fescue site, 1994 '
Icm
IOcm
100cm
IOOOcrn
10000cm
meof
0.017
a
0.035
a
0.082
b
0.355
C
2936
d
cema
0.010
a23
0.009
a
0.017
a
0.361
a
0.514
a
hel
0.087
a
0.128
a
0.229
a
8.492
a
agcr
0.003
a
0.004
a
0.008
a
0.001
a
brte
0.006
a
0.012
b
0.013
b
0.054
b
0.397
c
trac
0.028
a
0.024
a
0.020
a
0.172
a
3.752
b
zema
0.050
a
0.062
ab
0.073
ab
0.191
b
13.865
c
forbs
0.020
0.040
0.100
0.360
3.010
grass
0.020
0.040
0.030
0.160
2.070
total4
0.022
a
0.036
a
0.051
a
0.251
b
3.19
C
'Sites were not significantly different (p=0.05, Tukey's test).
-Letters indicate pairwise comparisons results (p=0.05). Compare across rows, and not down
columns.
1Kruskall -Wallace analysis of variance on ranks (knapweed. p=0.0250; sunflower, p=0.122; crested
wheatgrass. p=0.1534; wheat. p=0.0110) with Dunn's pairwise comparisons. One way analysis of
variance with (clover, pO.OOOl; cheatgrass, p=0.0001; com, p=0.0303) Student-Neuman-Kuels
pairwise comparisons used for com, clover, and cheatgrass. All weights for parametric anovas log
transformed.
'Total differences in plant weight from Tukey's Pairwise Comparisons. Weights were Log
transformed.
Table 24. Analysis o f variance, plant weights (log transformed), Idaho fescue site, 1994
Source
DF
Sum o f Squares
Mean Square
F Value
Pr>F
Model
10
139.76
13.98
43.12
0.0001
Error
21
6.81
0.32
Corrected
Total
31
146.56
Source
DF
Tvpe III SS
Mean Square
F Valui
Hole
4
91.09
22.77
70.26
0.0001
Spp
6
33.41
5.57
17.18
0.0001
R Square
C.V.
Root MSE
Log (wt) Mean
0.95
-22.14
0.57
-2.57
60
Table 25. Analysis o f variance on wheat and knapweed weight (log transformed) both
sites, 1993 and 1994._________________________
______
Source
DF
Sum o f Squares
Mean Square
F Value
Pr>F
Model
7
26 96
3 85
663
0.0001
Error
28
16.26
0 58
Corrected
Total
35
43.22
Source
DF
Type III SS
Mean Square
F Value
Pr>F
Hole
4
18.79
4.70
8 09
0.0002
Spp
I
3 52
3.52
6.07
0.0200
Site
I
2.81
2.81
4 83
0.0360
Year
I
1.24
1.24
2.13
0.1560
R Square
CV
Root MSE
Log (wt) Mean
0.623
-87.538
0.762
-0.871
Table 26: Effect o f gap size on wheat and knapweed weights (g): both sites, 1993 and
1994'
10000cm
I OOOcrn
100cm
IOcm
Icm
cema2"3
0.02
a4-5 0 02
a
0.03
a
0.31
b
1.19
b
trae
0.03
a
0.03
a
0 02
a
0.17
ab
2.74
b
total
0.02
a6
0.02
a
0 03
a
0.02
ab
2 66
b
'Sites significantly different (p=0.05), years are not.
-Species are significantly different (p=0.05).
3Ccma=knapweed. trae=whcat
'Letters indicate pairwise comparison results (p=0.05). Compare across rows, not down voiumns.
'Kruskall Wallace analysis of variance on ranks and Dunn's pairwise comparisons (p=0.05) used on
wheat and knapweed separately (wheat, p=0.00()I; knapweed. p=0.0001).
"Analysis of variance (p=0.0001) and Tukcy's test (p=0.05) used for pairwise comparisons used for
pooled species.
61
Table 27. Analysis o f variance on plant weights (log transformed), blue grama site, 1993
and 1994.
Source
DF
Sum o f Squares
Mean Square
F Value
Pr>F
Model
6
60.74
10.12
23 94
0.0001
Error
9
3.81
0.42
Corrected
Total
15
64.55
Source
DF
Type 111 SS
Mean Square
F Value
Pr>F
Hole
4
59 99
15.00
35.47
0.0001
Spp
I
0.73
0.73
1.72
0.2230
Year
I
2.08
2.08
4.93
0.0500
R Square
CV
Root MSE
Log (wt) Mean
0.94
-23.50
0.65
-2.77
Table 28. Effect of gap size on wheat and knapweed weights (g), blue grama site, 1993
and 1994.
______________________________________
Icm
10cm
100cm
I OOOcrn
10000cm
cema1
0 03
0.02
0.01
0.03
1.15
trae
0.04
0.01
total
0.02 a2'3
0.03 a
3 09
0.03 a
0.02 a
2.12b
lCema=Icnapweed, trae=wheat
-Letters indicate pairwise comparison results (p=0.05). Compare across rows, not down columns.
3Analysis o f variance (p=0.0001) and Tukcy's test (p=0.05) used for pairwise comparisons used for
pooled species.
62
Table 29. Analysis o f variance on wheat and knapweed weights (log transformed) at the
Idaho fescue grassland, 1993 and 1994_____________________________________
Source
DF
Sum of Squares
Mean Square
F Value
Pr>F
Model
6
89 26
14.88
3.30
0.03
Error
13
58.63
4.51
Corrected
Total
19
147.89
Source
DF
Txpc III SS
Mean Square
F Value
Pr>F
Hole
4
48 92
12.23
2.71
0.08
Spp
I
25 36
25 36
5 62
0.03
Year
I
14.98
14.98
3 32
0.09
R Square
CV
Root MSE
Log (wt) Mean
0.60
-152.26
2.12
-1.39
Table 30. Effect o f gap size on knapweed and wheat weights (g), Idaho fescue grassland.
1993 and 1994.
Icm
10cm
IOOcm
1000cm
10000cm
cema
0.01
0.02
0.04
0.43
1.15
trae
028
003
0.02
0.17
4.75
total
0.02 a
0.02 a
0.03 a
0.27 a
2.90 b
1Cema=Icnapweed, trae=wheat
-Letters indicate pairwise comparison results (p=0.05). Compare across rows, not down columns.
’Analysis of variance (p=0.0001) and Tukcy's test (p=0.05) used for pairw ise comparisons used for
pooled species.
63
Table 31. Relationship o f plant height (cm) to site, year, and hole size.
plant
log(ht)
trans
form1
Y
intercept
site
coef.
meof3
2
068
hean
I
agcr
year
coef.
log (hole)
coef.
log
(hole)2
coef.
R2
p value
-0.04
0.05
0.02
0.38
0.0001
3.99
-6.72
3 55
0.93
0.65
0.032
2
0.51
-0.04
-0.02
0.02
0.06
0698
bite
2
0.37
-0.09
0.18
-0.03
033
0.014
zema
2
0.78
-1.97
-1.54
1.43
0.26
0.005
trae93
I
3.31
-0.69
-1.60
1.09
0.31
0.0001
trae94
I
7.00
1.34
-4.08
1.33
0.33
0.0001
trac 93 &
94
2
683
0.82
-1.18
0 82
0.23
0.0001
cema93
I
0.46
-0.08
-0.11
0.07
0.49
0.0001
cema94
I
2.63
-0.30
-0.54
0.42
0.49
0.0001
cema 93
& 94
2
2.70
-0.65
-0.61
0.59
0.45
0.0001
-1.34
-0.73
'AU possible regression equations were tried, these fit the data best.
2I= no transformation. 2= log 10 transformation.
3Cabu=Shepard1S purse, meof=colver. cema=knapweed, hean=sunflower, agcr=crested wheatgrass.
brte=cheatgrass, trae=wheat, zema=com.
64
Table 32: Analysis o f variance on wheat and knapweed height data (log transformed) at
both sites, 1993.__________ ________________
Source
DF
Sum of Squares
Mean Square
F Value
Pr>F
Model
6
1.878
0.313
9 800
0.0003
Error
13
0.415
0.032
Corrected
Total
19
2.293
Source
DF
Type III SS
Mean Squa
F Value
Pr>F
Hole
4
‘ 1.196
0.299
9.360
0.0009
Spp
I
0.587
0.587
18.370
0.0009
Site
I
0.096
0.096
2 990
0.1072
R Square
CV
Root MSE
Log (wt) Mean
0.819
20.712
0.179
0.86
Table 33. Effects o f gap size on wheat and knapweed heights (cm) in both sites, 1993.
100cm
10cm
Icm
10000cm
I OOOcrn
cem a'2
3.26
a3-4 3.37
a
4.15
a
7.26
a
17.1
a
trae
6 88
a
9 28
a
10.2
a
8 85
a
29.80
b
total5
5.20
a
6.56
a
7.18
a
7.94
a
23.94
b
'Species are significantly different, sites are not.
2Cema=knapw eed. trae= wheat
’Letters indicate pairwise comparison results (p=0.05). Compare across rows, and not down
columns.
1Kruskall-Wallace analysis of variance on ranks (wheat. 0.0001: knapweed, p=0.0319) and Dunn's
pairwise comparisons used to analyze single species.
5 Anova (p=0.0003) and Tukcy's pairwise comparisons for pooled species (p=0.05).
65
Table 34. Analysis o f varianceon wheat and knapweed height (log transformed) at the
blue grama site, 1993,__________________________
Source
DF
Sum of Squares
Mean Square
F Value
Pr>F
Model
5
6.67
0.85
5.43
0.065
Error
4
0.62
3.28
21.03
0.010
Corrected
Total
9
730
Source
DF
Type III SS
Mean Sqm
F Value
Pr>F
Hole
4
3.39
085
5.43
0.07
Spp
I
3 28
3 28
21.03
0.0100
R Square
C.V.
Root MSE
Log (wt) Mean
0.91
21.63
0.40
1.83
Table 35: Analysis of variance on wheat and knapweed heights (log transformed) at the
Idaho fescue site, 1993._______________________________________________________
Source
DF
Sum of Squares
Mean Square
F Value
Pr>F
Model
5
4.20
0.84
22.00
0.005
Error
4
0.15
0.04
Corrected
Total
9
4.35
Source
DF
Type III SS
Mean Square
F Value
Pr>F
Hole
4
3.73
0.93
24.46
0.005
Spp
I
0.46
0.46
12.18
0.030
R Square
C.V.
Root MSE
Log (wt) Mean
0.96
9.11
0.20
2.15
66
Table 36. Effect of gap size on wheat and knapweed heights (cm) at the blue grama site,
1993.
Height
Icm
IOcm
100cm
IOOOcm
IOOOOcm
ccma'
3.48
2 80
3.48
2 80
12.7
trae
7.04
11.02
18.17
5.45
29.30
total
5.58 a:
7.06 a
7.45 a
3 ^
i
lCcma=Iaiapweed. trac=wheat
2Letters indicate pairwise comparison results (p=0.05). Anova (p=
comparisons (p=0.05)
21.00 a
id Tukey's pairwise
Table 37. Effect o f gap size on wheat and knapweed heights (cm) at the Idaho fescue site.
1993.
Height
Icm
10cm
100cm
I OOOcrn
10000cm
cema'
3.10
4.10
5 60
10.14
21.50
trae
6.73
7.40
8.42
10.55
30.30
total
4.87 a2
8 97 a
7.01 a
10.32 a
25.90 a
1Cema=Iaiapweed, trae=wheat
2Letters indicate pairwise comparison results (p=0.05). Anova (p=0.0()5). and Tukey's pairwise
comparisons (p=0.05).
67
Table 38. Effect o f gap size on plant heights (cm) in both sites, 1994.
Icm
IOcm
100cm
IOOOcrn
10000cm
meof'
5.38
a
5.98
a
7.77
a
13.00
b
28 26
C
cema
2.78
a23
2.72
a
3 38
a
6.90
b
7
b
hean
10.3
a
10.24
a
20.70
a
35.40
a
agcr
3.67
a
3.77
a
4.50
a
5.10
a
bite
3.18
a
4.95
a
5.35
a
4.00
a
5.73
a
trac
9.91
ab
24.40
ab
15.50
b
10.62
ab
8.34
a
zema
8 89
a
10.70
a
13.12
a
13.00
a
31.20
a
forbs
4.69
5 29
7.75
13.37
19.87
grasses
622
7.57
7.01
8.45
13.27
10.00
ab
16.40
7.37
b
a
a
6.52
a
5.51
total45
1Meof=clover, ccma=kn ipweed. he
-.flower, agcr=crcsted wheatgrass, brtc= cheatgrass,
trae=wheat, zema=com.
2Lctters indicate pairwise componsu.i results (p=0.05). Compare across rows, and not down
columns.
3One w ay analysis of variance w ith Student-Ncuman- Kuels pairw ise comparisons used on
sunflower (p= 0.2993). Kruskall-Wallace analysis of variance on ranks with Dunn’s pairwise
comparisons used on crested wheatgrass (p=0.7342), wheat (p=0.0051), knapweed (p=0.0002),
clover (p=0.0001), and cheatgrass (p=0.3719), and corn (p=0.2400).
“’Analysis o f variance (p=0.0001) and Tukey's Test used for multiple pairwise comparisons used for
pooled species and grass versus forb analysis.
5Spccies and sites were significantly different (p=0.05, Tukey's test).
68
Table 39. Effect o f gap size on plant heights (cm) at the blue grama site, 1994.
Height
Icm
meof
4.65
a
cema
2.08
al2-34 2.13
IOcm
bean
agcr
4 85
bite
trac
4.93
a
5.25
100cm
a
5.95
1000cm
a
10000cm
17.2
a
7.30
a
10.30
960
1.95
5.10
2.70
3 85
4.50
a
11.13
a
zema
forbs
4.04
5.05
grasses
4.90
3.73
total45
4.27
a
4.66
5.95
9.74
7.70
b
595
a
8.63
b
lCema=Iaiapweed, cema=knapweed, hean=sunflower, agcr=crested wheatgrass, brte=cheatgrass,
trae=xvheat, zema=com.
"Letters indicate pairwise comparison results (p=0.05). Compare across rows, and not down
columns.
2One way analysis o f variance with Student-Neuman- Kuels pairwise comparisons used on clover
(p=0.0197), knapweed (p=0.0512), and wheat (p=0.0338). Crested wheatgrass, com, sunflower,
and cheatgrass had too few responses to analyze.
3Analysis of variance (p=0.0400) and Tukey's test used for multiple pairw ise comparisons used for
pooled species and grass versus forb analysis.
4Species were significantly different (p=0.05, Tukey's test).
69
Table 40. Effect o f gap size on plant heights (cm) at the Idaho fescue site, 1994.
Icm
IOcm
IOOcm
1000cm
10000cm
meof
5.80
a
6.71
a
8.18
a
13.37
b
31.03
a
cema
2.97
a*3
296
a
338
a
6.90
a
7.18
a
bean
10.30
a
a
20.70
a
a
48.25
a
agcr
3.48
a
4.50
a
4.50
a
a
bite
3.18
a
5.40
b
5.35
a
4.00
a
668
b
trae
8.04
a
7.22
a
4.90
a
5 68
a
14.75
b
zema
8 89
a
11.37
a
13.12
a
12.98
a
31.23
a
forbs
4.96
5.44
7.96
13.37
24.93
grasses
636
8.23
7.01
8.45
13.27
total45
5.79
a
6.52
a
7.44
a
10.10
a
a
20.41
b
lMeof=Clover, ccma=knapvveed. hean=sunflower, agcr=crested wheatgrass, brte=cheatgrass,
trae=wheat, zema=com.
2Letters indicate pairwise comparison results (p=0.05). Compare across rows, and not down
columns.
3One way analysis o f variance with Student-Neuman- Kuels pairwise comparisons used on
sunflower (p=0.1669), crested wheatgrass (p=0.5437), cheatgrass (p=0.010), and wheat
(p=0.0035). Kruskall-Wallace anova on ranked data used for clover (p=0.0016), knapweed
(p=0.1099), and com (p=0.2046).
4Analysis o f variance (p=0.0001) and Tukey’s test used for multiple pairwise comparisons used for
pooled species and grass versus forb analysis.
5Species were significantly different (p=0.05, Tukey's test).
70
Source
DF
Sum o f Squares
Mean Square
F Value
Pr>F
Model
4
3.570
0.325
17.920
0.0001
Error
6
0.670
0.018
Corrected
Total
I
4.240
Source
DF
Type III SS
Mean Square
F Value
Pr>F
Hole
4
0993
0.248
13.720
0.0001
Spp
6
2.207
0.368
20.310
0.0001
Site
I
0.402
0.402
22.18
0.0001
R Square
CV
Root MSE
Log (ht) Mean
0.842
16.398
0.135
0.82
Table 42.
Analysis o f variance on plant heights (log transformed), blue grama site, 1994.
Source
DF
Sum of Squares
Mean Square
F Value
Pr>F
Model
9
4.93
0.55
3.67
0.04
Error
8
1.19
0.15
Corrected
Total
17
6.12
Source
DF
Type 111 SS
Mean Square
F Value
Pr>F
Hole
3
2.07
069
4.63
0.04
Spp
6
2.94
0.49
3.29
0.06
R Square
CV
Root MSE
Log (ht) Mean
0.81
23.69
0.39
1.63
71
Table 43. Analysis o f variance on plant height (log transformed), Idaho fescue site, 1994.
Source
DF
Sum o f Squares
Mean Square
F Value
Pr>F
Model
10
12.94
1.29
17.21
0.0001
Error
20
1.50
0.08
Corrected
Total
30
14.44
Source
DF
Type 111 SS
Mean Square
F Value
Pr>F
Hole
4
3.57
0.89
11.87
0.0001
Spp
6
9.28
1.55
20.57
0.0001
R Square
CV
Root MSE
Log (ht) Mean
0.90
13.44
0.27
2.04
72
Table 44. Effect o f gap size on wheat and knapweed heights (cm) at both sites, 1993 and
1994.'
Height
I
cm
cema2
3.10
a3-4
3.10
a
3.82
ab
7.2
ab
9.42
b
trae
7.09
a
8.57
a
7.39
a
7.41
a
16.89
b
total
5.25
a
5.92
a
5 78
a
7.30
a
13.16
b
10
cm
100
cm
1000
cm
10000
cm
'At p=0.05, Tukeys tests show species and years significantly different, but sites are not.
2Cema=knapweed, trae=wheat.
’Letters indicate pairwise comparison results (p=0.05). Compare across rows, not down columns.
4Single species are analyzed with Kruskall-Wallacc analysis o f variance on ranks and Dunn's
pairwise comparisons (knapweed. p=0.0018); wheat, p=0.0017). Pooled species analyzed with
analysis o f variance (p=0.0001)and Tukey's pairw ise comparisons (p=0.05).
Table 45. Analysis o f variance on wheat and knapweed height (log transformed), both
sites, 1993 and 1994.___________________________________________________
Source
DF
Sum o f Squares
Mean Square
F Value
Pr>F
Model
7
2.737
0.391
16.930
0.0001
Error
28
0.647
0.023
Corrected
Total
35
3.384
Source
DF
Type 111 SS
Mean Square
F Value
Pr>F
Hole
4
1.563
0.391
16.920
0.0001
Spp
I
0.846
0.846
36.630
0.0001
Site
I
0.179
0.179
7.730
0.0096
Year
I
0.287
0.287
12.450
0.0015
R Square
CV
Root MSE
Log (ht) Mean
0.809
19.064
0.152
0.797
73
Table 46. Analysis o f variance on wheat and knapweed heights (log transformed), blue
grama site, 1993 and 1 9 9 4 . _______________ _______________________________________
Source
DF
Sum o f Squares
Mean Square
F Value
Pr>F
Model
5
7.93
1.59
7.96
0.003
Error
10
1.99
0.20
Corrected
Total
15
9 92
Source
DF
Type III SS
Mean Square
F Value
Pr>F
Hole
4
4.16
1.04
5.22
0.020
Spp
I
3.77
3.77
18.90
0.001
R Square
CV
R oot MSE
Log (ht) Mean
0.80
26.20
0.45
1.70
Table 47. Analysis o f variance on wheat and knapweed height (log transformed) Idaho
fescue site, 1993 and 1994._________________________________________________________
Source
DF
Sum o f Squares
Mean Square
F Value
Pr>F
Model
5
5.48
1.10
7.58
0.001
Error
14
2.02
0.15
Corrected
Total
19
7.50
Source
DF
Type III SS
Mean Square
F Value
Pr>F
H ole
4
4.26
1.06
7.37
0.002
Spp
I
1.22
1.22
8.44
0.012
R Square
CV
Root MSE
Log (ht) Mean
0.73
19.58
0 38
1.94
74
Table 48. Effect o f gap size on wheat and knapweed heights (cm), blue grama site, 1993
and 1994'
Icm
IOcm
100cm
1000cm
10000cm
cema2
3.2
2.64
1.73
2.47
8 65
trae
6.72
10.59
13.17
5.45
15.68
total
5.22 a3
6.37 a
7.45 ab
3.66 a
12. 16b
'At p=().05, Tukcys tests show species and years significantly different.
2Ccma=knapwecd, Irae=WhcaL
'Letters indicate pairwise comparison results (p=0.05). Compare across rows, not down columns.
Table 49. Effect o f gap size on plant heights (cm), Ida io fescue site, 1993 and 1994 '
Height
Icm
10cm
100cm
1000cm
I 0000cm
cema2
3 05
3 48
4.39
9 23
10.04
trae
7.33
7.32
6.16
7.84
17.86
total
5.27 a2
5.60 a
5.38 a
8.44 ab
13.95 b
'At p=0.05, Tukeys tests show species and years significantly different.
2 Cema=knapw eed. trae=wheat
3Letters indicate pairwise comparison results (p=0.05).
75
Table 50. Precipitation at the blue grama site, compared to long term data.
T rid e n t
1993
p p t (cm )
Sep 92
Oct 92
Nov 92
Dec 92
Jan 93
Feb 93
Mar 93
Apr 93
May 93
Jun 93
Jul 93
Aug 93
Sep 93
2.90
3.20
1.04
0.58
0.97
0.91
0.66
6.20
4.70
8.36
14.15
7.01
2.92
53.59
43.33
Total:
Gr s e a . t o t .
T rid en t
1994
p p t (cm )
Sep 93
Oct 93
Nov 93
Dec 93
Jan 94
Feb 94
Mar 94
Apr 94
May 94
Jun 94
Jul 94
Aug 94
Sep 94
2.92
1.24
0.41
0.51
0.56
0.43
1.40
4.80
3.96
4.29
3.86
0.91
0.10
25.40
17.93
total
Gr s e a . tot.
m o n th ly lo n g term a v g
n o rm , (cm ) d e p a rt, (cm. s t d d e v
s td d e v
2 62
4.37
0.91
0.46
0.94
0.91
-0.41
9.88
3.84
11.23
25.12
11.02
2.18
73.07
63.27
-0.28
1.17
-0.13
-0.13
-0.13
0.00
-1.07
3.68
-0.86
2.87
10.97
4.01
-0.74
19.48
19.94
56.9
47.6
31.7
17.4
13.7
16.7
37.8
43.7
56.7
58.1
61.2
54.4
3.2
3.2
4.5
5.4
7.8
6.8
4.9
3.4
3.2
3.0
2.5
2.5
3.2
m o n th ly lo n g term a v g
s td d e v
d
e
p
a
rt,
(cm,
s td d ev
n o rm , (cm )
2.18
0.48
-0.61
0.36
0.13
0.15
1.07
7.09
2.36
3.10
4.55
-1.17
-3.45
16.23
12.47
-0.74
-0.76
-1.02
-0.15
-0.43
-0.28
-0.33
2.29
-1.60
-1.19
0.69
-2.08
-3.56
-9.17
-5.46
45.2
27.2
28.9
32.6
24.7
39.9
45.6
56.7
61.2
67
68.8
3.2
3.2
4.5
5.4
7.8
6.8
4.9
3.4
3.2
3.0
2.5
2.5
s td d ev
d e p a r tu r e
17.8
14.9
7.0
3.2
1.8
2.5
7.7
12.9
17.7
19.4
0.0
24.5
17.0
s td d ev
d e p a rtu re
14.1
8.5
6.4
6.0
3.2
5.9
9.3
16.7
19.1
26.8
27.5
76
Table 5 1. Precipitation at the Idaho fescue site compared to long term data.
B ozem an
Sep 92
Oct 92
Nov 92
Dec 92
Jan 93
Feb 93
Mar 93
Apr 93
May 93
Jun 93
Jul 93
Aug 93
Sep 93
Total:
Gr s e a . t o t .
B ozem an
Sep 93
Oct 93
Nov 93
Dec 93
Jan 94
Feb 94
Mar 94
Apr 94
May 94
Jun 94
Jul 94
Aug 94
Sep 94
total
Gr s e a . t o t
1993
p p t (cm )
5.46
5.66
2.69
2.03
1.17
0.61
1.57
8.08
8.41
10.74
12.57
7.34
3.07
69.42
50.22
1994
p p t (cm )
3.07
3.30
1.07
1.78
0.94
1.40
2.13
4.90
3.68
8.89
7.16
1.55
0.91
40.79
27.10
n o rm , (cm ) d e p a rt, (cm )
6.30
7.19
2.57
2.13
0.13
-0.48
-0.46
11.40
8.74
14.22
21.72
10.87
4.95
89.28
71.91
0.84
1.52
-0.13
0.10
-1.04
-1.09
-2.03
3.33
0.33
3.48
9.14
3.53
1.88
19.86
21.69
n o rm , (cm ) d e p a rt, (cm )
4.95
2.54
-0.89
1.68
-0.33
1.09
0.66
5.05
-0.71
10.52
10.90
-0.71
-3.12
31.62
21.92
1.88
-0.76
-1.96
-0.10
-1.27
-0.30
-1.47
0.15
-4.39
1.63
3.73
-2.26
-4.04
-9.17
-5.18
m o n th ly lo n g term a v g
s td dev
s td d ev
s td d ev
d ep a rtu re
56.7
49.7
31.2
18.8
18.0
19.4
38.2
43.5
56.3
57.4
58.6
60.7
55.2
3.5
3.1
4.5
4.7
6.7
6.0
4.9
3.6
3.2
3.1
2.5
25
3.5
16.2
16.0
6.9
4.0
2.7
3.2
7.8
12.1
17.6
18.5
23.4
24.3
15.8
s td d ev
lo n g term a v g
s td d ev
s td d ev
depa rtu re
55.2
46.1
27.8
28.7
32.6
24.8
39.7
46.1
56.2
61.1
67.5
69.2
61.4
3.5
3.1
45
4.7
6.7
6.0
4.9
3.6
3.2
3.1
2.5
2.5
3.5
15.8
14.9
6.2
6.1
4.9
4.1
8.1
12.8
17.6
19.7
27.0
27.7
17.5
77
Table 52. Percentages o f plants eaten by grasshoppers at both sites, 1993 and 1994
I cm
10 cm
100 cm
1000 cm
Im
total
bogr' 93
14
10
100
90
0
28
feid 93
22
17
60
50
0
26
bogr 94
I
3
4
4
0
2
feid 94
7
6
12
3
0
6
’Bogr= blue grama, feid Idaho fescue.
Table 53. Signed rank sum test results for grasshopper control, 1994 (Onsager and Henry
blue grama site
Idaho fescue site
In treated area
out o f treated area
In treated area
out o f treated area
no.1
I
11
0
6
p value2
0.03
0.01
'Number of grass ioppers counted.
2P value from Wilcoxson signed rank test (Ncter, Wasserman, and Kutner, 1990).
Table 54. Light intensity (pages) in gaps.
gap size
blue grama site
light intensity group'
Idaho fescue site
light intensity group
I cm2
a
a
10 cm2
a
a
100 cm2
a
a
1000 cm2
a
b
I m2
a
b
1Bonferroni pairw ise comparison
78
Table 55. Soi moisture (%), at both sites, 1994 '
May
July
September
gap size
feid site2-3 bogr site
feid site
bogr site
feid site
bogr site
I cm
13.95a4
1.66a
2l6ab
0.00a
11.28a
1.93a
10 cm
19.07a
2.50a
3.70a
1.35a
11.07a
2.23a
100 cm
44.34a
6.25a
2.27ab
1.04a
9.71a
1.53b
1000 cm
41.98a
6.43a
1.69ab
1.83b
7.57b
1.96a
I m
39.79a
4.84a
3.09a
1.13a
4.86b
1.83a
‘Idaho fescue soil moisture significantly higher than blue grama moisture (p=0.0001, MannWhitney rank sum test).
2Feid= Idaho fescue, bogr= blue grama site
3Kruskall-Wallace and Dunn's pairwise comparisons for feid May (p=0.267), Feid July (p=0.0367),
feid Sept. (p=0.0001), bogr July (p=0.0156), bogr Sept. (p=0.0003). Oneway anova with
Bonferroni pairwise comparisons used for bogr June (p=0.8171). 4Compare down rows, not across
columns.
79
Table 56. Emergence (%) o f eight species at both sites, 1993 and 1994.
1994 Idaho fescue
shepard s purse knapweed
12
76
I cm
33
80
10 cm
0
70
100 cm
0
60
1000 cm
0
100
1m
clover sunflower wheatqrass cheatqrass
84
80
76
84
80
90
87
87
90
90
40
50
100
90
30
70
100
75
50
100
wheat
100
93
100
90
100
com
1994 blue gramma site
Shepard's purse knapweed
92
52
1 cm
73
60
10 cm
10
30
100 cm
0
0
1000 cm
0
50
1m
clover sunflower wheatqrass cheatqrass
72
80
28
80
87
93
60
87
100
30
30
30
60
10
10
10
75
25
50
25
wheat
84
100
100
80
100
com
1993 Idaho fescue site
knapweed
92
I cm
80
10 cm
100
100 cm
100
1000 cm
1m
100
wheat
68
87
100
80
100
1993 blue qramma site
knapweed
60
1 cm
87
10 cm
100
100 cm
80
1000 cm
100
1m
wheat
72
93
100
100
100
96
100
70
80
100
88
80
30
0
0
80
Table 57. Survival (%) o f eight species at two sites, 1993 and 1994.
1994 Idaho fescue site
knapweed
1cm
60
10 cm
73
60
100 cm
1000 cm
20
1m
75
wheat
76
73
90
50
100
clover sunflower wheatgrass cheatqrass
88
8
44
52
60
53
33
33
90
20
40
20
90
0
30
0
100
50
0
100
com
1994 blue gramma site
knapweed
1cm
16
10 cm
27
100 cm
0
1000 cm
0
1m
75
wheat
16
7
0
0
75
clover sunflower wheatgrass cheatqrass
52
0
8
0
53
8
13
7
20
0
0
0
0
0
0
0
75
25
50
25
corn
1993 Idaho fescue site
knapweed
1cm
92
10 cm
67
100 cm
100
100
1000 cm
1m
100
wheat
88
87
100
80
100
1993 blue gramma site
knapweed
64
Icm
87
10 cm
60
100 cm
60
1000 cm
100
1m
wheat
92
93
60
40
100
64
93
50
60
75
shepard's purse
0
0
0
0
0
0
13
0
0
0
shepard's purse
0
0
0
0
0
81
Table 58. Knapweed and wheat heights at the blue grama site, 1993.
Triticum Aestivum
Centaurea maculosum
Gap Size (cm) Height (cm) Weight (g) Gap Size (cm) Height (cm) Weight (g)
7.5
0.00
I
1
5.0
0.05
1.7
0.00
1
0.8
1
0.01
0.02
1
1
7.3
3.4
0.02
0.00
1
2.3
1
3.2
0.01
1
0.01
0.8
2.9
1
0.01
0.00
1
1.2
1
2.1
0.01
5.7
0.00
3.7
1
1
0.02
8.7
0.01
1
7.0
0.07
1
6.8
0.02
1
0.02
1
3.9
0.01
1
0.04
2.9
4.2
1
10.5
0.03
1
0.03
1
3.7
7.1
0.00
0.07
1
1
4.6
0.01
4.9
0.02
1
1.6
1
0.05
7.1
0.02
1
5.5
1
0.02
0.00
1
10.0
3.2
1
0.02
3.0
10.3
0.02
1
1
10
4.2
0.02
11.8
0.01
1
0.04
10
1.4
0.02
10.8
1
0.01
10
4.0
5.9
0.01
1
0.04
0.7
0.00
10
9.1
1
0.02
0.00
10
1.9
8.4
1
0.02
4.8
10
10.2
0.01
1
0.04
3.5
10
8.7
0.01
1
0.03
4.8
10
0.02
9.1
10
0.01
1.3
0.04
10
8.7
10
2.9
0.01
10
0.01
11.5
10
0.01
1.4
10
0.02
12.2
10
4.9
0.02
10
0.01
9.2
10
0.01
0.6
10
0.02
11.9
10
0.02
2.7
100
0.02
14.2
10
0.00
0.9
100
0.02
9.2
10
0.01
1.6
100
0.01
11.9
10
0.04
3.4
1000
0.01
14.2
10
0.01
1.8
1000
0.02
3.1
10
0.04
1000
2.2
0.01
11.2
10
1.73
12.7
10000
0.03
11.3
10
0.01
16.6
10
0.02
7.7
100
0.04
16.5
100
0.08
15.3
100
0.13
6.8
1000
0.03
4.1
1000
8.75
29 3
10000
82
Table 59. Knapweed and wheat heights and weights at the Idaho fescue site, 1993.
Triticum Aestivum__________________ Centaurea maculosum_______________
Gap Size (cm) Height (cm) Weight (g) Gap Size (cm) Height (cm) Weight (g)
1
8.9
0.01
1
1.0
0.00
1
7.3
0.01
1
1.6
0.00
1
13.1
0.01
1
2.9
0.01
0.01
1
1
9.0
3.8
0.01
1
9.4
0.00
1
2.3
0.01
1
4.4
0.00
1
2.2
0.02
5.4
1
0.01
1
4.8
0.02
1
9.9
0.01
1
3.0
0.01
I
10.9
0.01
1
2.0
0.01
7.4
1
0.00
1
0.01
1.3
1
7.0
0.02
1
3.4
0.01
0.02
1
1.7
0.01
3.3
1
1
0.01
0.01
9.1
2.8
1
7.5
0.05
0.03
1
1
5.3
0.04
0.02
1
1
3.8
4.0
0.01
8.6
0.02
4.4
1
1
4.4
0.05
1
0.02
1
4.3
3.4
0.04
1
0.02
1
5.5
2.9
0.02
3.4
0.02
1
1
3.4
0.02
1
0.03
1
2.1
0.01
0.03
1
3.1
1
1.2
4.0
0.02
1
7.9
0.02
1
0.01
0.02
1
2.3
10
6.0
0.02
0.02
10
5.1
5.6
10
0.05
10
8.3
0.11
10
13.1
0.01
10
2.3
9.4
0.01
10
0.01
10
1.4
0.02
10
9.2
3.4
0.02
10
0.01
10
2.1
0.00
1.1
0.02
10
3.5
10
0.04
0.04
10
6.5
13.2
10
0.04
5.8
0.04
10
10
11.0
3.5
0.03
10
0.03
6.9
10
0.03
10
3.6
0.04
2.0
10
0.26
100
10.0
0.04
10.7
10
0.03
4.1
100
0.01
3.5
10
0.04
7.0
100
0.05
14.4
100
0.03
4.9
100
0.08
4.6
100
2.0
0.00
100
0.03
2.0
100
1.12
13.0
1000
0.01
8.3
100
0.24
10.4
1000
0.01
12.8
100
0.14
9.5
1000
0.13
3.5
1000
0.13
6.2
1000
0.02
1000
14.0
0.64
11.6
1000
0.01
19.1
1000
4.68
21.5
10000
0.01
5.6
1000
7 86
30.3
10000
83
Table 60. Grass heights and weights at the blue grama site, 1994.
B r o m u s te c to ru m
A g r o p y ro n c ris ta tu m
H eight
G a p S iz e (cm ) H eig h t (cm ) W e ig h t (g) G a p S iz e
2.7
3.6
0 .0 0 9
10
I
5.1
6.1
10000
1
0 .0 1 3
26
0 004
10000
10
2 .3
10
1.6
0 .0 0 3
10000
5.1
0.001
W e ig h t
0 .0 0 3
0 .1 6 6
0014
T riticum A e s tiv u m
G a p S iz e
H eight
1
26
I
4.6
I
4 .3
I
8.2
10
45
10000
14.5
10000
9 .5
W e ig h t
0 .0 1 3
0 .0 3 2
0 .0 2 8
0 .0 5 4
0 .0 2 4
1 .2 1 6
0 .201
Zeam ays
G a p S iz e
10
10
H eight
4.2
7.1
W e ig h t
0 .0 6 2
0 .0 8 2
84
Table 61. Grass heights and weights at the Idaho fescue site, 1994.
A a ro p y ro n c h s ta tum
B ro m u s te cto ru m
TrWcum A e stivu m
Z ea m a ys
IGap Size (cm) Height (cm) Weight (q) Gap Size Height Weight Gap Size Height Weight Gap Size Height Weight
5.9
I
I
0.004
2.1 0.001
I
I
4.8 0.015
8.7 0026
3.9
1
3.2 0.001
1
0.003
I
4.5 0.022
1
8.1 0.083
I
I
2.6
4.1 0.004
I
0.002
I
7.6 0.023
10.7 0.044
I
3.5
3.7 0.005
I
I
0.003
I
12.2 0.052
9.1 0.038
1
8.1 0042
I
1.9
0.000
1
18.3 0.033
I
14.7 0.060
4.1 0.002
I
1
3.9 0.016
I
2.6
0.005
I
13.8 0.063
I
3.3 0.001
I
34
6.3 0.021
I
I
0.003
15.2 0.129
1.7 0002
1
I
6.2 0.025
0.001
1
6.2 0.039
1
3.5
I
1.0 0.001
I
7.6 0.037
0.005
I
14.3 0.073
I
2.5
I
1.6 0.001
I
10.1 0.086
I
3.2 0.010
3.2
0.003
I
1
2.1 0.006
2.1
0.003
1
7.8 0.016
I
3.3 0.046
I
5.0
0.003
10
6.1 0.008
I
7.1 0.026
I
6.1 0.083
I
86
10
7.6 0.019
I
9.9 0.039
I
5.5 0.023
10
0.001
0.024
1
9.7 0.027
2.4
4.8 0.009
8.1
0.000
10
I
10
I
6.7 0.040
9.2 0.018
4.3
0.007
10
3.8 0.010
I
10
I
8.3 0.016
6.9 0.022
100
I
3.6
0.008
4.3 0.008
10
10
8.6 0.072
1
6.2 0.007
100
4.3 0005
3.6
0.003
10
0.040
10 14.6 0.102
1
92
100
4.7 0.020
6.5
0.010
100
5.4 0.020
10
I
8.6 0.034
100
8.1 0.018
0.005
100
2.5
14.1
0.048
10
4.2 0.007
10
8.1 0.020
1000
10
16.0 0.068
3.7 0.101
10
2.1 0.004
1000
10 16.2 0.074
3.0 0.013
10000
6.2 0.343
10
10 13.1 0.121
10
4.1
0.021
10000
5.1 0.260
10 12.3 0.074
10
9.2 0.023
10000
6.5 0.138
12.4 0.074
10
8.9 0.848
10
9.2 0.022
10000
10
9.7 0.064
7.0 0.047
10
10 13.2 0.049
10
8.8 0.034
10
8.6 0.033
10
7.2 0.010
10
6.9 0.031
0.012
10
59
10
10.9 0.053
10
8.8 0.029
100
14.2 0.070
100
6.0 0.023
100
16.6 0.065
100
3 3 0.016
100
14.3 0.043
4.3 0.024
100
7.4 0.085
100
100
4.1 0.015
100
13.1 0.083
100
3.2 0.009
1000
15.0 0.178
5.6 0.012
100
1000 25.0 0.670
100
5.1 0.052
1000
14.2 0.040
100
6.3 0.017
1000
6.0 0.046
100
6.2 0.012
1000
6.1 0.140
1000
5.2 0115
1000
11.6 0.072
4.6 0.043
1000
10000 67.8 13.865
3.5 0.009
1000
10000
19.8 0.722
4.9 0.060
1000
10000
6.1 0.125
1000
10.2 0.641
0.033
10000
5.1
3.753
10000
191
10000
19.6 8.250
10000
15.2 2.972
85
Table 62. Forb heights and weights at the blue grama site, 1994.
Melilotus officionales
Centaurea maculosum
Gap Size (cm) Height (cm) Weight (g) Gap Size
1
3.4
0.007
1
1
0.025
5.3
1
I
0.004
4
1
1
0.017
46
1
0.005
10
1
3.2
0.014
10
I
5.9
0.033
10
1
5.5
0.045
10
1
7.4
* 0.019
10000
4.3
1
4.9
0.069
10000
I
5.9
0.025
1
0.008
3.9
1
0.001
1
2.1
0.038
4.6
10
0.03
5.3
10
0.043
5
10
0.043
7.2
10
0.079
7
10
0.012
4.5
10
0.031
4.7
10
0.012
3.7
10
0.082
7.6
100
0.026
43
100
1.518
17.2
10000
Height
2.8
1.9
1.1
2.5
1.9
0.9
3.1
2.6
10.2
9.2
Weight
0.017
0.01
0.003
0.018
0.009
0.004
0.02
0.018
1.052
0.654
Helianthus annus
Gap Size Height Weight
10
10.5
0.153
10
9.7
0.185
10000
9.6
1.801
86
Table 63. Forb heights and weights at the Idaho fescue site, 1994.
M e lilo tu s o ffic io n a le s
Gap Size (cm) Height (cm) Weight (g)
0.7
0000
I
6.4
0.021
I
5.7
0.013
1
0.005
4.5
1
0.021
I
6.7
0.002
1
4.0
0.001
I
3.2
I
0.006
3.9
1
5.2
0.011
4.0
0.006
1
6.2
0.011
I
8.1
0.023
1
0.018
8.2
1
48
0 008
1
4.4
0.006
I
10.3
0.046
1
9.2
0.019
1
9.6
0.041
I
0.002
1
1.4
0.022
1
5.2
7.3
0 060
1
6.8
0.029
I
0.034
7.5
10
4.1
0006
10
0.012
5.1
10
4.4
0.032
10
0.013
9.1
10
0.087
6.1
10
0 068
7.7
10
0.024
98
10
0 095
7.4
100
0.068
14.2
100
0 006
9.0
100
0.080
4.1
100
0.015
4.2
100
0.066
6.1
100
0.037
78
100
0.041
7.1
100
0.333
8.2
100
0.229
12.9
1000
0.596
12.2
1000
0434
13.7
1000
0 175
17.7
1000
0.454
10.1
1000
0.312
9.9
1000
0258
11.0
1000
0 046
12.7
1000
0.695
5.7
1000
0 367
27.3
10000
0 362
10.4
10000
6.208
59.3
10000
4.807
51.5
10000
C e n ta u re a m a c u lo s u m
Gap Size
1
1
1
1
1
1
I
I
1
I
I
I
1
I
I
10
10
10
10
10
10
10
10
10
10
100
100
100
100
100
100
10000
10000
10000
10000
Height
2.5
2.5
3.7
4.1
2.9
3.1
1.3
1.8
2.1
3.9
2.3
4.2
3.2
4.1
2.8
1.0
3.4
2.6
3.1
2.7
15
4.0
4.7
2.3
4.3
6.0
0.4
2.4
4.1
4.2
3.2
5.7
2.7
10.2
10.1
H e lia n th u s a n n u s
Weight
0.005
0.034
0.012
0.009
0.013
0.005
0.001
0.000
0.010
0.011
0.011
0.011
0.007
0.006
0.006
0.004
0.010
0.006
0.006
0.002
0.016
0.021
0.005
0.014
0.032
0.003
0.001
0.008
0.022
0.031
0.022
0.614
0638
0.007
0.117
Gap Size
1
1
10
10
10
100
100
10000
10000
Height
7.3
13.3
19.9
9.8
1.3
23.1
18.3
41 6
54.9
Weight
0.007
0.166
0276
0.073
0.036
0.252
0.205
12.100
4.883
87
Y=0.613651 ♦3i2BECGX
RSquered = 0227
--------
RagreaGcn
............
95%Q
—-—
se%R
Figure I . Effect o f gap size on forb weight (log scale).
Figure 2. Effect o f gap size on grass weight (log scale).
88
Y=-I SCB08 + O.123B79K
RSquared = 0.510
—
rtegwgcon
.......
95%a
Figure 3. Effect o f gap size on forb height (log scale).
1
I
:
..... _________
.
C- 0 ~
J=
=0
I
I
“
30
......
.
>
Y=-190271 8 15E02X
RSovered =0296
t
-3 —
4
• •
•
—
I
0
l
i
5
-----
Ragreasm
............
S6%a
l
10
15
Log(gap)2
Figure 4. Effect o f gap size on grass height (log scale).
89
1 0O O
10000
Log(gap)
Figure 5. Clover weights at both sites. Weights and gaps are log transformed. Hollow
boxes are responses from the blue grama site. Solid boxes are responses from
the Idaho fescue site. These are average values.
90
1000
10000
Log(gap)
Figure 6. Sunflower weight at both sites. Weights and gaps are log transformed. Hollow
boxes are responses from the blue grama site. Solid boxes are responses from
the Idaho fescue site. These are average values.
91
100
1000
10000
Log(gap)
Figure 7. Crested wheatgrass weight at both sites. Weights and gaps are log transformed.
Hollow boxes are responses from the blue grama site. Solid boxes are
responses from the Idaho fescue site. These are average values.
92
O
-1
10O
1000
10000
Log(gap)
Figure 8. Cheatgrass weights at both sites. Weights and gaps are log transformed.
Hollow boxes are responses from the blue grama site. Solid boxes are
responses from the Idaho fescue site These are average values.
93
q
-1
1 0O O
10000
Log(gap)
Figure 9. Corn weights at both sites. Weights and gaps are log transformed. Hollow
boxes are responses from the blue grama site. Solid boxes are responses from
the Idaho fescue site. These are average values
94
q
-1
10O
1000
10000
Log(gap)
Figure 10. Knapweed weights at both sites, 1993. Weights and gaps are log transformed.
Hollow boxes are responses from the blue grama site. Solid boxes are
responses from the Idaho fescue site. These are average values.
95
1000
10000
Iog(Gap)
Figure 11. Knapweed weights at both sites, 1994. Weights and gaps are log transformed.
Hollow boxes are responses from the blue grama site. Solid boxes are
responses from the Idaho fescue site. These are average values.
96
O
-1
10O
1000
10000
Log(gap)
Figure 12. Wheat weights at both sites, 1993. Weights and gaps are log transformed.
Hollow boxes are responses from the blue grama site. Solid boxes are
responses from the Idaho fescue site. These are average values.
97
q
-1
1000
10000
Log(gap)
Figure 13. Wheat weights at both sites, 1994. Weights and gaps are log transformed.
Hollow boxes are responses from the blue grama site. Solid boxes are
responses from the Idaho fescue site. These are average values.