This file was created by scanning the printed publication. Errors identified by the software have been corrected; however, some errors may remain. Geographic Fariationm PonderosaPine Leader Growth BY JAMES W. HANOVER Abstract.Growthof the shootaplcesof 91 treesin a 4S-year-old Pinusponderosa Laws.provenance testwasmeasured periodlcally with a transit.Analysis of the measurements led to the followingconclusions: (1) 19 racesof ponderosa pineplanted nearPriestRiver, Idaho,showed phenologica], morphological, or physiological variationin sixcharacters: dateo.fbeginning growthin the spring,dateof endinggrowth, totalseasonal elongation, duratlo3q of growth,lengthof dormantterminalbud,and rapidityof growth;(2) thevariation described is continuous, although it wasnot possible to asse.ciate an environmental gradient to thevariation pattern;(3) all trees except thelocalKaniksu source achieved theirmaximum surgeof growthduringthe same7-dayperiod,May 20-27,andthisresponse closely paralleled temperature fluctuation; (4) sim?le correlations among the19progenies between totalseasonal elongation and eitherbeginning date,relativerapidity,or durationof growthwerenot significant. Correlation between dormant terminal budlengthandtotalseasonal elongationwasstrong.Also., budlengthwaspositively correlated withrelative rapiditywhich wasnegatively correlated with duration of growth. FEW CRITICAL OBSERVATIONS have been made of intraspecific variationin the seasonalactivityof shootgrowthon tall trees. A 45-year-oldprovenance testof ponderosap•ne(Pinus•3onderosa Laws.) provided a good sourceof material for measuring growth of mature tree apiceswith some environmentalcontrol The objectivexvas to determinewhetherprovenances of ponderosapineshowvariationin date of beginningor cessation of leadergrowth, rationof growth,rapidityof growth,total seasonal elongation, andlengthof dormant (1936) and Morris et al. (1957) observedtimesof lateralbudbursting on 18and 40-year-old Douglas-fir provenance plantations.They suggestthat this characteristicis under stronggeneticcontrol. I•ens-Moller (1957) attemptedto determinethe degreeof geneticcontrolover time of bud burstingby measuringit on about 700 Douglas-firseedlings,repre- sentingsevenelevations from 60 to 4,000 feet. The measurements were made after the seedlings had overwinteredin Corvallis,Oregon. His resultsindicatedsig- terminal bud. Measurements indicate that nificant differences betxveen certain sources each of thesegrowth characteristics vary significantly withinthe species. The variation is closelyassociated with seedsource in dateof budburstingthe followingspring location. Pest Studies of Shoot Growth Bud burstingof Douglas-fir,Pseudotsuga menzlesli(Mirb.) Franco, has received considerable attention.Mungerand Morris 86 / ForestScience and "some correlation" b.etween this char- acteristicand elevationof seedlingsource. Chingand Bever (1960) observed dates of bud bursting of Douglas-fir nursery The authoris a geneticist,Intermonntain Forestand RangeExpt. Sta., ForestService, U.S. Dept.Agric.,andisstationed at Moscow, Idaho. seedlings representing 14 sourcesvarying in latitude(42ø20' to 50ø30') andelevation(100 to 4,100 feet). They found close relation between latitude or altitude of sourceand date .of bud bursting,althoughthe southernmost sourcewas significantlyearlierthan other p.rovenances. Vaartaja (1954, 1959) hasshownpatterns of geographicvariationin growth responsefor many tree species.Kriebel (1957) obtainedevidenceof clinal geographic variationin dates when sugar maple(ricer saccharum Marsh.) from 18 seedsources commenced growth. IrgensMoller (1958) demonstrated geographic variationin time of height growth cessation amongDouglas-firseedlings.Critch- field (1957), workingwith lodgepole pine (Pinus contortaDougl.) seedlings, associated a continuousvariation in growth durationwith changes in altitudeand latitudeof seedsource.Langlet (1959) analyzed Wright and Baldwin's (1957) data on the growthrate of 46 Scotchpine (Pinussylvestris L.) provenances. He interpretedthegeographic variability in height incrementas continuous and relatedto day lengthandtemperature. Baldwin(1956) measuredterminal growth during two seasons on 9- and 10-year-oldtreesof a Europeanlatch (Larix declduaMill.) provenance plantation in New Hampshire but found no correlation between either •ltitude or latitude of seed source and all speciesoccurredabout the sametime. Tryon and Finn (1937) notedthiswhen theyobservedgrowthcurvesfor red pine ( Pinusrexinosa Ait.), whitespruce( Picea glceuca(Moench) Voss), and European larchin New York State.Friesner(1942) notedthesimilarityin datesof peakgrowth within treesof Pinusbankdanaand P. sylvestrisL. Fowells(1941) comparedseasonalheightgrowthof sixconifersgrowing within a 10-acre area in California and re- porteddifferencesin beginning,duration, and rapidityof growth. This brief reviewof pastwork indicates clearly that the genetic makeup.of the source material and a number of climatic factorsmay interactto producegeographic variationin shootgrowth characteristics. Through naturalselection, populations becomeadaptedto each local environment. Sometimes the adaptation canbestatistically correlated with environmental factors. In othercasesthisis not poss•lebecause interpopulationmigration, limited populati.on sizes,and slow geneticresponse to selection pressuretend to obscurephenotype correlations.Under growthchamberconditionsor by appropriatesamplingtechniquesand experimentaldesign,the influence exerted on growth by any one environmental factormightbedetermined; but this would be difficult for the species and growth characteristics being considered here. •m.ountof currentheightgrowth or mid- pointsof growthcurves.He observed that: (1) all sources completed50 percentof their growth at approximatelythe same date,(2) beginning dateof growthvaried as greatlywithin as betweenprovenances, and (3) air temperaturewas associated with bud bursting. Kienh.olz (1941), Brown (1915), Friesnet (1942), McMillan ( 1957), andothershaveattributed the inceptionof bud growth largely to temperature effects within certain photo- periodic andgenetic limitations. Severalauthorshave comparedterminal growth betweenspeciesgrowing at one 2tocality.The peakof seasonal growth for Experimental Materials Seedor seedlings representing 22 provenancesof ponderosa pine were plantedat the Priest River ExperimentalForest in northern Idaho from 1911 to 1917. Weidman (1939) hasfully described the climatenear the pointof originof eachof the 22 seedsources and hasalsogenerally described the geographic areasof the seed sources(see also Table 1 and Fig. 1). The seedsources, or progenies, fairly well representthe broad range of ponderosa pine in the westernUnited States.Since detailsof plot establishment were well devolume9, hum,herl, 1963 / 87 T.4BLE 1. Geographic locationof ponderosa pine seedsources represented at Priest River Experimental Forest. • General geographicarea and national forest Elevation Latitude Longitude Feet North Pacific Siskiyou 2,000 42 ø 05' 123ø 40' Boise 5,500 43 ø 30' 115ø 00• Payette 5,000 44-0 30' 116 North Plateau Whitman 5,000 44 ø 38' 118ø Umatilla 3,500 46ø 00' 117ø 25' 30• Bitterroot Bitterroot Bitterroot 4.,000 5,000 7,200 46 ø 46 ø 46 ø 00' 00' 00' 114ø 114ø 114ø 20' 20' 20' Lolo Kanlksu Colville 3,000 2,600 2,700 47ø 48 ø 48 ø 10' 20' 40' 114ø 116ø 119ø •0' 50' 00' 7,100 35ø 10' Ill ø 50' 8,000 35ø 40' 105ø 30' Roosevelt Ashley (Central Plateau) Custer 8,000 7,500 3,200 40ø 40ø 45ø 30' 40' 30' 105ø 109ø 104ø 40' 40' 00'• 50' South Plateau Coconlno Santa Fe East of Continental Divide Helena 4,500 46 ø 30' 111o Harney (Black Hills) 5,000 43ø 40' 103ø 30' San Isabel 8,000 38ø 00' 105ø 00•' •Elevation 2•380feetl latitude48ø20'• longitude116050 '. scribedby Weldman they are not elaborated here. Originally each progenyplot was composed of either 100 or 50 trees, but mortalityof varyingdegreeshas reduced the number of trees remaining to 10 to 50 per plot. All treesfrom one of the originalsources(Mt. Shasta,California) were eliminatedentirelyby a severe temperaturedrop in December 1924 (Kempff 1928). Two sources of questionableoriginalsowere omittedfrom this study. The Priest River test of seed sources has receivedconsiderable studysinceits establishment. Individual progenyhave been previously examined and compared for differences in vigor and frost resistance (Kempff 1928); external and internal foliagecharacteristics (Weldman 1939); total heightand diametergrowth (Weldman 1939, Squillaceand Silen 1962); and seasonof cambialgrowth (Daubenmire 1950). 88 / ForestScience Methods and Definitions Methods of measure'men:t. Seasonalactivity of shootapiceshas never been describedfor these progenies. In March 1958 five dominant or codomlnant trees rangingfrom 25 to 65 ft in heightwere selected for measurement from each of the 19 plots.Their initialterminalbudlength and subsequent heightgrowthwere measuredthroughout the course of thegrowing season. The convenient method of direct meas- urementabovea constantpoint, suchas a pininserted at thebase.oftheterminalbud• was impractical because of the tree height and top flexibility. Therefore, the indirect method of angular measurementswith a transitwasadopted.Nine fixedinstrument stations were established from which the tipsand basesof all selected treescouldbe sightedclearly. The distancefrom each tree to a station was measured. Graduated sticks were nailed to the base of each tree to permitadjustmentfor heightof instru-- ment at each remeasurement. The onlydifiqculty encountered in using this method was causedby wind. Any shghtbreezeconsiderably reducedthe accuracyof an observation l consequently, all measurements weremadeonlyduringcalm periodsof the day, usuallyearly morning or evening. Error in measurement result•ng from the exper/mentaltechniquedid not exceed --+0.03 ft, well within the precisionnecessary for statisticalvalidity. This value was determinedby averaging I FmURE1. Climaticregions of ponderosa pinerange(northof Mexico):(I) NorthPacific; (2) northplateau;(3) centralplateau;(4) southplateau;(5) eastof Coutin,ental Divide; (6) SouthPacific.The brokenline shows therangeof ponderosa pineaccording to Sudworth (1917) andMunns(1938). Localities fromwhichseedusedin experiment wasderivedareindicated by blackdots.The experimental siteis located at the Kaniksu dot. (From Weldman19.39) volume9, numberI, 1963 / 89 the variationbetweenreadingsobtained to theseheightswere usedin the analyses. from successive measurements duringthe Mirov et al. (1952) useda method based on lines of intersection of observed periodwhennogrowthwasoccurr/ng.It includes errordueto instrument resetting, weekly increment values and average residual treeleanaftera wind,andhuman weeklyincrementover the entiregrowing error. Sincethe tipsof four treesdiedduring the observation period,completemeasurements were obtainedon only 91 trees. Treeswere measured every2 to 4 days beginning April 2. Fr.omMay 13, when all treeswere growing,until July 8, measurements were taken about every 7 days.Thereafter, untilOctober 4, longer period. Others have suggested obtaining the beginningand endingdatesby constructinggrowth curvesfor eachtree and determining thedesired pointson thecurve. Dependingon the natureof the data,any of thesemethods may be appropr/ate. Duration of growth is the number of dayselapsingbetweenthe 5 percentand 95 percentdates.Total elonga'tion is total intervals between measurements sufficiently increment in feet between the beginning accountedfor later growth and new bud and endingbasedates.Bud lengthin feet formation. Definitions.The beginning and ending datesof growthweredetermined bycomputingthe pointsof time (expressed as consecutive daysfrom January1) at which 5 percentand 95 percentof the totalseasonalgrowthhad beencompleted.This was necessary because of the difficultyin pin-pointing actualgrowthbeginning and cessation. Early growthseems to besporadicat firstbecause of dailyweathervariationsand because smallperiodicincrements may beobscured by error in measurement. Calculation of thesebasedates(.or 0 percent and 100 percent)is necessary to arrive at the 5 percentand 95 percent points.The beginning basedate height was taken as the averageof the first and all successive measurements up to one showinga difference of greaterthan 0.05 ft betweentwo consecutive readingsfollowedby no successive readings differing by lessthan 0.05 ft. Ending basedate heightwasdetermined by takingthe average of successive measurements not differing by more than 0.10 ft after encountering a difference of less than 0.10 ft between two successive measurements. Rather than 0.05, 0.10 ft was used to adjust for the increasedtime betweenobservationsat the closeof growth. Once the beginningand endingbasedateswere was computedfrom transitreadingsto the baseand tip of the budson April 1, when the first heightmeasurements were made. Rapidity of growth, as usedhere, is the percentof the total season's growth completed within the 2-week periodof most rapidgrowth. Rapiditymay be thoughtof in two ways:(1) absolute, whichis total elongation per unit of time, and (2) relative, which eliminates differences in total elongation andcompares growthrate during the grand periodof seasonal growth. The latterismoreappropriate in this study becausedifferent geneticpopulations have differentabsolute growthrates(Tables2 and 3). Absoluterapiditymay be more applicablefor most intraracial or intraspecificstudies,sincelessgeneticdiversity isusuallyencountered. The periodof most rapid growth for all the ponderosa pine progenies represented in thisstudywas the 2-weekperiodin which38 to.58 percent of their total growth .occurred.Fowells (1941) considered relativerapidityin h•s studyof seasonal growthof youngponderosapinein California. However, hismethods of calculation were different. He used a constantpercentof growth (50 percent) rather than a constanttime interval, and he measured rapiditybytheminimumnum- ber of days requiredto completeth•s amountof growth. found,the5 percentand95 percent heights Method of analysis.From the data obwerecalculated.The datescorresponding tained,mean valuesfor eachprovenance 90 / ForestScience TABLE 2. Ranking ofponderosa pinesources bydateof beginning andending ofgrowth and durationof growthfi Mean beginningdate Source Mean endingdate Day of year• Kaniksu 114 Ashley 115 Slsklyou 116 Bitterroot7,200' 117 Umatilla Source Mean duration Day of year2 Bitterroot7,200' Siskiyou 153 155 117 Lolo Helena Boise 156 156 157 Bitterroot4,000' 118' Bitterroot4,000' Roosevelt San Isabel Colville Lolo Helena 119 119 119 120 120 Payette Harney 120 120 Santa Fe 120 Bitterroot 5,000' 121 Boise. Custer 121 122 Ashley Whitman Coconlno 123 124 Santa Fe 169! Coconino 175J Source Days Boise Helena Lolo 35 36 37 Bitterroot7,200' 37 Custer 37 158 Bitterroot5,000' 38 Kaniksu Colville Custer 158 158 159 Colville 39 Siskiyou 40 Whitman 40 Bitterroot 5,000' 159 Bitterroot4,000' 40 San Isabel Roosevelt 160 161 Payette Payette Harney 161 162 Whitman 163 Umatilla 164 167'} 41 42 San Isab Harney Roosevelt Kanlksu Santa Fe Umatilla Ashley Coconino aAnytwomeans notincluded withinthesameline appearing to the rightof eachrankingof sources aresig- n•ficantly different. aDaysof the yearnumbered consecutively from Januu'yI, 1998. TABLE 3. Rankingof ponderosa pinesources by totalelongation, terminalbudlength (April 1958), andrelativerapidityof growth) Mean total elongation Sourc.e Harney Mean terminalbud length Feet 0.93 Roosevelt .98 Rapidity Source Feet Source Roosevelt 0.11 Santa Fe Coconlno Umatilla 37.9 42.3 43.5 Ashley .I 2 Percent Payette 1.06 Coconino .13 Custer 1.08 Colville .14 Bitterroot7,200' 1.11 Custer .14 Payette Harney 43.7 44.6 San Isabel 1.16 San Isabel .14 Colville 1.22' I Harney .15 Bitterroot5,000' Bitterroot4,000' Ashley 45.3 45.6 46.2 Helena Boise S•sklyou 1.23! 1.23/ 1.24/ Payette Bitterroot7,200' Bitterroot5,000' .15 .16 .16 Lolo 47.1 B•tterroot 4,000' 1.28/ Bitterroot4,000' .18 Coconlno Kaniksu 1.29[ 1.33I Siskiyou Kaniksu .20 Ashley 1.12 ] Santa Fe.14 .21 Sisklyou 47.3 San Isabel Whitman 47.6 48.0 Custer Roosevelt Colville 48.6) 48.7/ 50.3/ Bitterroot 7,200' Bitterroot 5,000' 1.37 [ Umatilla .231 Whitman 1.43[ Helena Lolo 1.55[ Whitman Umatilla 1.64• Lolo .23 .24 .24 Boise Kaniksu Helena 51.4/ 51.9/ 55.0/ 58.1J •Anytwomeansnot included •vithinthe samellne appearing to the rightof eachrankingof sources are slgmficantly different. were calculated for the traits observed (Tables 2 and 3). Either individualtree data, progenymeans,or both were used for analysisof varianceof eachcharacteristic (Table 4). Bartlett's (1937) chisquaretest for homogeneity of variancein volume9, number1, 1963 / 91 TABLE 4. Analyscs of variance of characteristicsof leader growth among 19 sources of ponderosa p•ne. Growth characterlstic t F-value Beginning date Ending date 2.09* 2.80** Duration 3.05** Total elongation 1.97' Bud length prior to growth Rapidity (relative) 3.54** 2.39** XBasedon 5 percentand 95 percentpointsof seasonalgrowth completion. *p <0.05• *•p <0.01. vided further analysisof the variation amongthe 19 progenies.Any two means (Tables2 and 3) not includedwithinthe sameline appearingat the right of each setof dataare significantly different. For instance,in beginningdate of growth the Siskiyousourceis significantlydifferent from the Custer but not from the Boise. The data were alsoanalyzedfor simple correlations between measured attributes (table 5). Results and Discussion TABLE 5. Coefficients of simple correlation (r) between shootcharacteristics of ponderosa pine sources at PriestRiver. Character•stlcs Progenyor seedsourcebasis(n : Bud length and total elongation Duration and total elongation r Beginningdate and total elongatlon t Bud length and relativerapidity Duration and relative rapidity r 19) 0.724*** 0.133 0.008 0.458* --0.599** Beginning date and total heightof progeny t --0.135 Indlv•dualtree basis(n • 91) Bud length and total elongation 0.405*•'* Relativerapidityandtotal elongation --0.126 Bud lengthand relativerapidity 0.257* XBased on 5 percent and 95 percent of seasonal growth completion. *p <0.05• **p <0.01• *** p <0.001. progeny means showed that differences amongindividuals withinprovenances were nonslgnificant at the 5-percent level. This wastruefor all characteristics exceptend•ng dateand durationof growth,which were nonsignificant at the 1-percentlevel. That is, the five individual values from eachpr.ovenance wereconsistent enough to justifyuseof a single statistic forthesource. In thetwoexceptions, theindividuals varied too muchto permitassumption that they werefrom the samepopulations; but they The analyses of varianceshowsignificant differences amongprovenances for all the growth characteristics studied. Multiple range testsshow that no two adjacently ranked provenancesdiffer significantly, when arrayedaccordingto magnitudeof response, in any of the traits, i.e., under the conditionsof this study, the variation is continuous. The resultsof simplecorrelationanalyses betweenattributesprovidesomeindication as to the association between these re- sponses and total seasonalgrowth. At leastduring the year of this study,neither beginningdate, relativerapidity,duration of growth, nor ending date appearedto have any relationto total growth during the year. A strong positivecorrelation existedbetweenbudlengthand total elongation. Bud lengthwas alsosignificantly correlatedwith relativerapidityof growth. Thus, total elongatlonseemedto depend more on inherent growth potential than on an endogenous rhythmmechanism that gave certainindividualsor progeniesan advantageover poorerperformingones. Another interestingpoint revealedby simplecorrelationanalysiswas a highly significant negative correlation between relativerapidityand durationof growth. An effort was made to relate the con- tinuous variation patterns to geographic have some value when aberrant individuals location of the seed sources. W'eidman were removed. (1939) and Squillace and Silen (1962) established relationsbetweenmorphology or height growth of the ponderosapine Applying Kramer's(1956) extension of Duncan's(1955) multiplerangetestpro92 / ForestScience progeniesat Priest River and certain cli- in the upperportionsof the total elongation scale,and among thesethe sources The manyinteractingfactorsinvolvedand from higherelevations begangrowthlater. the inadequacy of the originalplot design Sources representing geographic regionsin makeit difficultto associate the responses which September-through-June precipitastudied here with seed source climatic tion is relatively low, generally began factors. growth later and grew less than seed Pictorial representation of severalclisourcesfrom areaswhere precipitation is matic factorsand growth responses (Fig. high for thisperiod. It is interesting that 2) provides insight intocauses of thevariathe local Kaniksusourcebegangrowth tion. This methodhasbeenappliedeffecconsiderably earlier than the two most southerlysources,Santa Fe and Coconino. tivelyin taxonomic studies .ofracesor speo c]eswhen multipleattributesare involved Four sources--Ashley, Umatilla,SantaFe, (Anderson1956). However,evenby this and Coconino--seem to have a distinctly methodonly certain trendsare evidenced longerperiodof growththanothersources. and no statistical significance is implied. Suchbroadgeneralizations aboutprogeny For instance,the easternsourcestend to performancein relatl.onto seedsourceare all that these data allow. groupinto a sectorof the diagramhaving a low totalelongation and a late beginning The period at which each tree and date. Only the north plateausources are progenyachievedtheir maximumsurgeof matic factors at the seed source localities. 1.70 I.B0 1.50 1,40 .8O 114 April 24 115 25 116 26 117 27 118 28 119 29 120 50 121 May I 122 2 125 5 124 4 Beginning Date of Grawth (Days numbered consecutively from Jan. l,1958) FIGURE2. Pictoriallze•d scatterdiagramshowlnff variationin threegrowthfactorsand threeclimatic factorsfor ponderosa pineseedsources. volume9, num3er1, I963 / 93 TdBLE 6. Weatherat PriestRiver Experimental Forestin relationto growthof ponderosapine sources. Temperatures Date Cloud Maximum Minimum øF øF Precipitation cover x Inches Week precedingmaximumgrowth: May 13 May 14 May 15 May 16 May 17 May 18 May 19 63 74 76 78 79 85 82 30 29 35 36 35 33 43 Tr. 0 0 0 0 0 0 7 2 76.7 34.4 0 2.4 20 21 22 23 24 25 26 85 88 88 85 86 85 90 44 55 46 47 55 51 50 0 0 0 0 0.01 0 0 Mean 86.7 49.7 0.00 91 89 73 76 73 69 72 47 52 46 41 42 41 38 0 Tr. 0 0 0.19 0.20 0 77.6 43.9 0.06 Mean 0 1 7 0 0 Week of maximum growth: May May May May May May May Week following maximum growth: May 27 May 28 May 29 May 30 May 3l June l June 2 Mean 0 2 1 10 2 4 5 3,4 1 6 9 5 10 4 6 5.9 •0 z cleari 10 z overcast. growth was related to local temperature. With few exceptions the period of most rapid growth for all trees was the same 7-dayperiod,May 20-27. Only oneprogeny, the local Kaniksusource,averaged1 response or for any of the othersconsideredhere. However,Munger and Moriis ined. Temperaturefluctuationcloselyparalleled this growth response.The week of maximumgrowth coincidedwith day and night temperatures higher than those of the weeks immediatelyprecedingand immediatelyfollowingthe week of maxi- (1936), Friesnet (1942), Tryon and Finn (1937), and Kienholz (1941) showedthat thoughyearlyvariationsoccur in the absolutevaluesfor manyphenologlcal characteristics,different individuals, races,or species usuallymaintainthe same relativepositions duringsuccessive years. All treesbegangrowingin the spring beforedangerfrom frost had passedand ceasedgrowing before the beginningof fall frosts. Also, growth was at least95 percent completedbefore maximum day lengthsand the July-Augustdry period were reached. This growth pattern is mumgrowth(Table 6). The temperature characteristicof many coniferous trees pattern of this year could have obscured whoseshootelongationdepends partly on storedcarbohydrates but alsoon cell elon- week earlier. In an effort to relate an environmentalfactor to this response, weather records at the Priest River Ex- perimentalForestheadquarters (lessthan •-mile from the plantation)were exam- some differences between sources for this 94 / ForestScience 12:307-310. gationand divisionin budsformed dur/ng the previous year (Sacher1954). KRIEBEL,HOWARDB. Literature Agr. Exp. Sta. Res.Bull. 791. 56 pp. LANGLET,OLOV. 1959. A cline or not a Cited ANDERSON, EDGAR.1956. Characterassociation analysis asa tool for the plant breeder. Brookhaven Symposia in BiologyNo. 9:123140. B^LDWIN,HENRY I. 1956. The perio.dof height growth in different provenances of Europeanlatch. F.A.O. Document56/4/ B^RTLE•r, M. S. 1937. Someexamplesof statistical methodsof research in agriculture and applied biologyßJ. Roy. Star. Soc. (Suppl.)4:137-183. BROWN,H. P. 1915. Growth studiesin forest trees,Pinus stro.bus.Bot. Gaz. 59:197- Genet. 8:13-22. McM1LLAN,CALVIN. 1957. Nature of the plant community.IV. Phenological variation within five woc>dlandcommunities un- der controlledtemperatures. Am. J. Bot. 44:154-163. LIDDICO•T. 1952. Altitudinal races of Pinusponderosa--a12-yearprogress report. J. For. 50:825-831. MORRIS, W•LL1AMG., R. R. S1LEN,and H. IRG•Ns-MoLLER. 1957. Consistencyof bud burstingin Douglas-fir.J. For. 55: 208-210. 241. CUING, KIM K., and DALE BEVER. 1960. Provenoncestudy of Douglas-firin the PacificNorthwestRegion. I. Nurseryperformange. Silvae Genet. 9:11-17. CroTCHYIELD, WILL•AMB. 1957. Geographic variation in Pinus contorta. Maria Moors Cabot Foundß,Harvard Univ. Publ. 3. 118 PP. DAURENmRZ,R. F. 1950. A comparison of seasonof cambialgrowth in different geographic racesof Pinusponderosa. Bot. 112:182-188. MUNOZR,THORNTON t., and W•LLIAMG. MORRIS. 1936. Growth of Douglas-fir trees of known seed source. Forest Service, U.S. Dept. Agric.Tech. Bull. 537. 40 pp. MUNNS, E. N. 1938. The distributionof importantforesttreesof the United States. Forest Service,U.S.. Dept. Agric. Misc. Publ. 287. 176 SACHER, J. A. 1954. Structure andseasonal activityof the shootaplcesof Pinuslambertianaand Pinusponderosa. Amer. J. Botany 41:749-759. DUtYCAN,D. B. 1955. Multiple range and multiple F tests. Biometrics11:I-42. FOWELLS, H. A. 1941. The period of seasonalgrowth of ponderosapine and associatedspecies.J. For. 39:601-608. FRIESNER,RAy C. 1942. Some aspectsof tree growth. Proc. Indiana Acad. Scl. 52: 36-44. SQUILLACE, A. E., and RoY R. SILEN. 1962. Racialvariationin ponderosa pine. Forest ScienceMonograph2. 27 •.p.. SVDWORTH, G•ORCmB. 1917. The pine treesof the RockyMountain regionßForest Service,U.S. Dept. Agric. Bull. 460. 47 pp. TRYON,H. H., and R. F. FINN. 1937. Notes IRGENs-MoLLER, HELGE. 1957. Ecotypicresponseto temperatureand photoperiodin Douglas-fir.For. Scl. 3:79-83. ß 1958. Genotypic variation in time of cessation of heightgrowthin Douglas-fir. cllne--a questionof Scotspine. Silvae MiRo% N. t., J. w. DUVVIELD,and A. R. 2556. 8 pp. Gaz. 1957. Patternsof geneticvariationin sugarmaple. Ohio For. Sci. 4:325-330. on the terminalgrowthof coniferous plantations in the Hudson highlands. Black RockForestPapersI(9) :54-56. VAARTAJA, O. 1954. Photoperiodlc ecotypes of trees. Canad.J. Bot. 32:392-399. 1959. Evidence of photo- KEMPvv, GERHARD.1928. Non-indigeno.us periodicecotypes in trees. Ecol. Mono- westernyellowpine plantations in northern graphs29:91-11 I. W•m•IAN, R. H. 1939. Evidencesof racial influencein a 25-year test of ponderosa Idahoß Northwest Sci. 2:54-58. KIEN•OLZ, RAYMOND.1941. Seasonalcourse of height growth in somehardwoodsin Connecticut.Ecology22:249-258. KRAMER, CLYDE YOVNG. 1956. Extension of multiple rangeteststo gro.upmeanswith unequalnumbersof replications.Biometrics pineßJ. Agric. Res.59:855-887. WRIGHT, J. w., and HENRY I. BALDWIN. 1957. The 1938 International Unic>n Scc>tch pine prc>venance testin New Hampshire. Silvae Genet. 6:1-14. volume9, number1, 1963