Structures in the Parameter Planes Dynamics of the family of complex maps n F (z) z zn with: Paul Blanchard Toni Garijo Matt Holzer U. Hoomiforgot Dan Look Sebastian Marotta Monica Moreno Rocha Elizabeth Russell Yakov Shapiro David Uminsky First a brief advertisement: AIMS Conference on Dynamical Systems, Differential Equations and Applications Dresden University of Technology Dresden, Germany May 25-28 2010 Organizers: Janina Kotus, Xavier Jarque, me One half hour slots for speakers. Interested in attending/speaking? Contact me at bob@bu.edu Structures in the Parameter Planes Dynamics of the family of complex maps n F (z) z zn What you see in the dynamical plane often reappears (enchantingly so) in the parameter plane.... Cantor Necklaces: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. A Cantor necklace in a Julia set when n = 2 Cantor Necklaces: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. A Cantor necklace in a Julia set when n = 2 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. and in the parameter plane Cantor Necklaces: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. A Cantor necklace in a Julia set when n = 2 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. and in the parameter plane Cantor Necklaces: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. A Cantor necklace in a Julia set when n = 2 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. and in the parameter plane Mandelpinski Necklaces: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Circles of preimages of the trap door and critical points around 0 Mandelpinski Necklaces: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Circles of preimages of the trap door and critical points around 0 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Circles through centers of Sierpinski holes and baby M -sets in the param-plane Mandelpinski Necklaces: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Circles of pre-preimages of the trap door and pre-critical points around 0 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Circles through centers of Sierpinski holes and baby M*-sets in the param-plane * the only exception Mandelpinski Necklaces: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Circles of pre-preimages of the trap door and pre-critical points around 0 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Circles through centers of Sierpinski holes and baby M*-sets in the param-plane * the only exception Mandelpinski Necklaces: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Circles of pre-preimages of the trap door and pre-critical points around 0 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Circles through centers of Sierpinski holes and baby M -sets in the param-plane Mandelpinski Necklaces: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Circles of pre-preimages of the trap door and pre-critical points around 0 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Circles through centers of Sierpinski holes and baby M -sets in the param-plane Mandelpinski Necklaces: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Circles of pre-preimages of the trap door and pre-critical points around 0 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Circles through centers of Sierpinski holes and baby M -sets in the param-plane Mandelpinski Necklaces: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Circles of pre-preimages of the trap door and pre-critical points around 0 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Circles through Sierpinski holes and baby Mandelbrot sets in the parameter plane As Douady often said “You sow the seeds in the dynamical plane and reap the harvest in the parameter plane.” It is often easy to prove something in the dynamical plane, but harder to reproduce it in the parameter plane. Here is how we will do this: Suppose you have some object in the dynamical plane that varies analytically with the parameter maybe a closed curve, maybe a Cantor necklace, or: dynamical plane Suppose you have some object in the dynamical plane that varies analytically with the parameter maybe a closed curve, maybe a Cantor necklace, or: Maybe it’s your face, so call it Face( ) dynamical plane Suppose you have some object in the dynamical plane that varies analytically with the parameter maybe a closed curve, maybe a Cantor necklace, or: Maybe it’s your face, so call it Face( ) Change , and Face( ) moves analytically: dynamical plane Suppose you have some object in the dynamical plane that varies analytically with the parameter maybe a closed curve, maybe a Cantor necklace, or: Maybe it’s your face, so call it Face( ) Change , and Face( ) moves analytically: maybe like this dynamical plane Suppose you have some object in the dynamical plane that varies analytically with the parameter maybe a closed curve, maybe a Cantor necklace, or: Maybe it’s your face, so call it Face( ) Change , and Face( ) moves analytically: or like this (you’re so cute!) dynamical plane So any particular point in Face( ), say the tip of your nose, nose( ), varies analytically with nose( ) dynamical plane So any particular point in Face( ), say the tip of your nose, nose( ), varies analytically with nose( ) dynamical plane So any particular point in Face( ), say the tip of your nose, nose( ), varies analytically with nose( ) dynamical plane So we have an analytic function nose( ) from parameter plane to the dynamical plane nose( ) parameter plane dynamical plane So we have an analytic function nose( ) from parameter plane to the dynamical plane nose( ) parameter plane dynamical plane Now suppose we have another analytic function G( ) taking parameter plane to dynamical plane one-to-one nose( ) G parameter plane dynamical plane So we have an inverse map G-1 taking the dynamical plane back to the parameter plane nose( ) G-1 parameter plane dynamical plane Now suppose G takes a compact disk D in the parameter plane to a disk in dynamical plane, and nose( ) is always contained strictly inside G(D) when D . nose( ) D parameter plane G-1 G(D) dynamical plane So G-1(nose( )) maps D strictly inside itself. nose( ) D parameter plane G-1 G(D) dynamical plane So G-1(nose( )) maps D strictly inside itself. So by the Schwarz Lemma, there is a unique fixed point * for the map G-1(nose( )). nose( * ) * D parameter plane G-1 G(D) dynamical plane * is the unique parameter for which G( * ) = nose( * ). * nose( ) * D parameter plane G-1 G(D) dynamical plane * is the unique parameter for which G( * ) = nose( * ). If we do this for each point in Face( ), we then get the same “object” in the parameter plane. G(D) D parameter plane dynamical plane * is the unique parameter for which G( * ) = nose( * ). If we do this for each point in Face( ), we then get the same “object” in the parameter plane. G(D) D Why are you so unhappy living in the parameter plane? parameter plane dynamical plane The goal today is to show the existence in the parameter plane of: 1. 2. 3. 4. Cantor necklaces Cantor webs Mandelpinski necklaces Cantor sets of circles of Sierpinski curve Julia sets 1. Cantor Necklaces A Cantor necklace is the Cantor middle thirds set with open disks replacing the removed intervals. 1. Cantor Necklaces QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. A Cantor necklace is the Cantor middle thirds set with open disks replacing the removed intervals. a Julia set with n = 2 and a Cantor necklace 1. Cantor Necklaces QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. A Cantor necklace is the Cantor middle thirds set with open disks replacing the removed intervals. a Julia set with n = 2 and another Cantor necklace 1. Cantor Necklaces QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. A Cantor necklace is the Cantor middle thirds set with open disks replacing the removed intervals. a Julia set with n = 2 and lots of Cantor necklaces 1. Cantor Necklaces QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n=2 And there are Cantor necklaces in the parameter planes. 1. Cantor Necklaces QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n=2 1. Cantor Necklaces QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n=2 We’ll just show the existence of this Cantor necklace along the negative real axis when n = 2. Recall: B = immediate basin of T = trap door F (z) z 3 z3 B QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. T Recall: B = immediate basin of T = trap door 2n free critical points c 1/2n F (z) z 3 z3 B c QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. T Recall: B = immediate basin of T = trap door 2n free critical points c 1/2n F (z) z 3 v 2 z3 B c Only 2 critical values v QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. T Recall: B = immediate basin of T = trap door 2n free critical points c 1/2n Only 2 critical values v 2 2n prepoles p ( )1/2n F (z) z 3 z3 B p c v QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. 0 p Consider F (z) z 2 z 2 , 0 Consider F (z) z 2 z 2 , 0 Since 0 , F preserves the real line. Consider F (z) z 2 z 2 , 0 need a glass of wine??? graph of F (x) Consider F (z) z 2 z 2 , 0 B = basin of infinity graph of F (x) Consider F (z) z 2 z 2 , 0 B = basin of infinity T = trap door graph of F (x) Consider F (z) z 2 z 2 , 0 B = basin of infinity T = trap door The two intervals I0 and I1 are expanded over the union of these intervals and the trap door. I1 I0 graph of F (x) Consider F (z) z 2 z 2 , 0 B = basin of infinity T = trap door The two intervals I0 and I1 are expanded over the union of these intervals and the trap door. So there is an invariant Cantor set on the negative real axis. I1 I0 graph of F (x) Consider F (z) z 2 z 2 , 0 B = basin of infinity T = trap door The two intervals I0 and I1 are expanded over the union of these intervals and the trap door. So there is an invariant Cantor set on the negative real axis. Add in the preimages of T to get the Cantor necklace in the dynamical plane for 0 . I1 I0 graph of F (x) QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. The Cantor necklace for negative QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. This portion is also a Cantor necklace lying on the negative real axis for 0 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. And we have a similar Cantor necklace lying on the negative real axis in the parameter plane for n = 2. To see this, let D be the half-disk |z| < 1, Re(z) < 0. D To see this, let D be the half-disk |z| < 1, Re(z) < 0. Let G( ) be the second iterate of the critical point G() F (v ) D To see this, let D be the half-disk |z| < 1, Re(z) < 0. Let G( ) be the second iterate of the critical point G( ) F (v ) (2 ) (2 )2 2 D To see this, let D be the half-disk |z| < 1, Re(z) < 0. Let G( ) be the second iterate of the critical point 1 G( ) F (v ) (2 ) 4 2 4 (2 ) 2 D To see this, let D be the half-disk |z| < 1, Re(z) < 0. Let G( ) be the second iterate of the critical point 1 G( ) F (v ) (2 ) 4 2 4 (2 ) 2 So G is 1-to-1 on D, and maps D over itself; G D -3.75 .25 To see this, let D be the half-disk |z| < 1, Re(z) < 0. Let G( ) be the second iterate of the critical point 1 G( ) F (v ) (2 ) 4 2 4 (2 ) 2 So G is 1-to-1 on D, and maps D over itself; equivalently, G-1 contracts G(D) inside itself. G-1 D -3.75 .25 For any in D (not just 0), we also have an invariant Cantor set as we showed earlier: For any in D (not just 0), we also have an invariant Cantor set as we showed earlier: U and U are portions 0 2 of a prepole sector U2 c U0 For any in D (not just 0), we also have an invariant Cantor set as we showed earlier: U and U are portions 0 2 of a prepole sector that are each mapped univalently over both U0 and U2. U2 c U0 For any in D (not just 0), we also have an invariant Cantor set as we showed earlier: U and U are portions 0 2 of a prepole sector that are each mapped univalently over both U0 and U2. So there is a portion of a Cantor set lying in U2. c U0 For any in D (not just 0), we also have an invariant Cantor set as we showed earlier: U and U are portions 0 2 of a prepole sector that are each mapped univalently over both U0 and U2. So there is a portion of a Cantor set lying in U2. And we can add in the appropriate preimages of the trap door to get a Cantor necklace. c U0 For any in D (not just 0), we also have an invariant Cantor set as we showed earlier: U and U are portions 0 2 of a prepole sector that are each mapped univalently over both U0 and U2. So there is a portion of a Cantor set lying in U2. And we can add in the appropriate preimages of the trap door to get a Cantor necklace. G(D) And, since lies in D, the Cantor set lies inside G(D). We can identify each point in the Cantor set in U2 by a unique sequence of 0’s and 2’s: s = (2 s1 s2 s3 ....) given by the itinerary of the point. U2 U0 We can identify each point in the Cantor set in U2 by a unique sequence of 0’s and 2’s: s = (2 s1 s2 s3 ....) given by the itinerary of the point. So, for each such sequence s, we have a map zs (), which is defined on D and depends analytically on zs ( ) U2 U0 We therefore have two maps defined on D: G(D) D We therefore have two maps defined on D: 1 1. The univalent map G( ) 4 4 G D G(D) We therefore have two maps defined on D: 1 1. The univalent map G( ) 4 4 2. The point in the Cantor set zs ( ) G D zs G(D) 1 So G ozs maps D strictly inside itself; G-1 D zs G(D) 1 So G ozs maps D strictly inside itself; by the Schwarz Lemma, there is a unique fixed point *s in D for this map. G-1 *s D zs G(D) 1 So G ozs maps D strictly inside itself; by the Schwarz Lemma, there is a unique fixed point *s in D for this map. * * G( ) z ( For this parameter, we have s s s ), so this is the unique parameter for which the critical orbit lands on the point zs ( ) . G-1 *s D zs G(D) * This produces a Cantor set of parameters s , one for each sequence s. Claim: this Cantor set lies on the negative real axis. * This produces a Cantor set of parameters s , one for each sequence s. Claim: this Cantor set lies on the negative real axis. Recall: G() 4 1/ 4 , so G decreases from .25 to -3.75 as goes from 0 to -1 in D. * This produces a Cantor set of parameters s , one for each sequence s. Claim: this Cantor set lies on the negative real axis. Recall: G() 4 1/ 4 , so G decreases from .25 to -3.75 as goes from 0 to -1 in D. the Cantor set in the dynamical plane lies on the negative real axis when 0. * This produces a Cantor set of parameters s , one for each sequence s. Claim: this Cantor set lies on the negative real axis. Recall: G() 4 1/ 4 , so G decreases from .25 to -3.75 as goes from 0 to -1 in D. the Cantor set in the dynamical plane lies on the negative real axis when 0. So G( ) must hit each point in the Cantor set along the negative axis at least once. * This produces a Cantor set of parameters s , one for each sequence s. Claim: this Cantor set lies on the negative real axis. Recall: G() 4 1/ 4 , so G decreases from .25 to -3.75 as goes from 0 to -1 in D. the Cantor set in the dynamical plane lies on the negative real axis when 0. So G( ) must hit each point in the Cantor set along the negative axis at least once. So each parameter in the parameter plane necklace must also lie in [-1, 0]. This produces the Cantor set portion of the necklace on the negative real axis. * s Similar arguments produce parameters on the negative axis that land after a specified itinerary on a particular point in B (that is determined by the Böttcher coordinate). Similar arguments produce parameters on the negative axis that land after a specified itinerary on a particular point in B (that is determined by the Böttcher coordinate). And then these intervals can be expanded to get the Sierpinski holes in the necklace. QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. 2. Cantor webs Recall that, when n > 2, we have Cantor “webs” in the dynamical plane: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n=3 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n=4 2. Cantor webs Recall that, when n > 2, we have Cantor “webs” in the dynamical plane: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n=3 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n=4 2. Cantor webs Recall that, when n > 2, we have Cantor “webs” in the dynamical plane: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n=3 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n=3 2. Cantor webs When n > 2, we also have Cantor “webs” in the parameter plane: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n=3 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n=4 2. Cantor webs When n > 2, we also have Cantor “webs” in the parameter plane: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n=3 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n=4 2. Cantor webs When n > 2, we also have Cantor “webs” in the parameter plane: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n=3 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n=4 A slightly different argument as in the case of the Cantor necklaces works here. Say n = 3. In the dynamical plane, we had the disks Uj. U1 v U2 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. v U4 n=3 U5 A slightly different argument as in the case of the Cantor necklaces works here. Say n = 3. In the dynamical plane, we had the disks Uj. U1 U0 v U2 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. v U4 U3 n=3 U5 Each of these Uj were mapped univalently over all the others, excluding U0 and Un, so we found an invariant Cantor set in these regions. A slightly different argument as in the case of the Cantor necklaces works here. Say n = 3. In the dynamical plane, we had the disks Uj. U1 U0 v U2 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. v U4 U3 n=3 U5 Each of these Uj were mapped univalently over all the others, excluding U0 and U3, so we found an invariant Cantor set in these regions. U0 and U3 are mapped univalently over these Uj, so there is a preimage of this Cantor set in both U0 and U3 Now let G( ) be one of the two critical values, so G( ) 2 And choose a disk D in one of the “symmetry sectors” in the parameter plane: D QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. U2 v U3 U4 Now let G( ) be one of the two critical values, so G( ) 2 And choose a disk D in one of the “symmetry sectors” in the parameter plane: Then G maps D univalently over all of U0, so we again get a copy of the Cantor set in D G U1 U0 v D QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. U2 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. v U3 U4 U5 Now let G( ) be one of the two critical values, so G( ) 2 And choose a disk D in one of the “symmetry sectors” in the parameter plane: Then G maps D univalently over all of U0, so we again get a copy of the Cantor set in D G U1 U0 D QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. v U2 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Then adjoining the appropriateSierpinski holes v gives a Cantor web in the parameter plane. U4 U 3 U5 3. “Mandelpinski” necklaces A Mandlepinski necklace is a simple closed curve in the parameter plane that passes alternately through a certain number of centers of baby M-sets and the same number of centers of S-holes. 3. “Mandelpinski” necklaces A Mandlepinski necklace is a simple closed curve in the parameter plane that passes alternately through a certain number of centers of baby M-sets and the same number of centers of S-holes. oops, sorry.... 3. “Mandelpinski” necklaces A Mandlepinski necklace is a simple closed curve in the parameter plane that passes alternately through a certain number of centers of baby M-sets and the same number of centers of Sierpinski-holes. 3. “Mandelpinski” necklaces QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. parameter plane n = 4 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. A Julia set 3. “Mandelpinski” necklaces QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. parameter plane n = 4 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. There is a “ring” around T passing through 8 = 2*4 preimages of T 3. “Mandelpinski” necklaces QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. parameter plane n = 4 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. There is a “ring” around T passing through 8 = 2*4 preimages of T 3. “Mandelpinski” necklaces QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. parameter plane n = 4 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Another “ring” around T passing through 32 = 2*42 preimages of T 3. “Mandelpinski” necklaces QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. parameter plane n = 4 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Another “ring” around T passing through 32 = 2*42 preimages of T 3. “Mandelpinski” necklaces QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. parameter plane n = 4 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Another “ring” around T passing through 128 = 2*43 preimages of T Now look around the McMullen domain in the parameter plane: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. parameter plane for n = 4 Now look around the McMullen domain in the parameter plane: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. There is a ring around M that passes alternately through the centers of 3 = 2*40 + 1 Sierpinski holes and 3 Mandelbrot sets Now look around the McMullen domain in the parameter plane: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. There is a ring around M that passes alternately through the centers of 3 = 2*40 + 1 Sierpinski holes and 3 Mandelbrot sets Now look around the McMullen domain in the parameter plane: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Another ring around M that passes alternately through the centers of 9 = 2*41 + 1 Sierpinski holes and 9 “Mandelbrot sets”* *well, 3 period 2 bulbs Now look around the McMullen domain in the parameter plane: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Another ring around M that passes alternately through the centers of 9 = 2*41 + 1 Sierpinski holes and 9 “Mandelbrot sets”* *well, 3 period 2 bulbs Now look around the McMullen domain in the parameter plane: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Then 33 = 2*42 + 1 Sierpinski holes and 33 Mandelbrot sets Now look around the McMullen domain in the parameter plane: QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Then 129 = 2*43 + 1 Sierpinski holes and 129 Mandelbrot sets Similar kinds of rings occur in the other parameter planes: n=3 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. parameter plane for n = 3 Similar kinds of rings occur in the other parameter planes: n=3 S0 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. S0: 2 = 1*30 + 1 Sierpinski holes & M-sets n=3 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. S1: 4 = 1*31 + 1 Sierpinski holes & M-sets* *well, two period 2 bulbs n=3 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. S2: 10 = 1*32 + 1 Sierpinski holes & “M-sets” n=3 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. S3: 28 = 1*33 + 1 Sierpinski holes & M-sets n=3 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. 82, 244, then 730 Sierpinski holes... n=3 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. sorry, I forgot..... nevermind the 13th ring passes through 1,594,324 Sierpinski holes... Theorem: For each n > 2, the McMullen domain is surrounded by infinitely many simple closed curves Sk (“Mandelpinski” necklaces) having the property that: 1. each Sk surrounds the McMullen domain and Sk+1, and the Sk accumulate on the boundary of M; 2. Sk meets the center of exactly (n-2)nk-1 + 1 Sierpinski holes, each with escape time k + 2; 3. Sk also passes through the centers of the same number of baby Mandelbrot sets* * with one exception The critical points and prepoles all lie on the 1/2n “critical circle” r | | c p p c The critical points and prepoles all lie on the 1/2n “critical circle” r | | The critical circle is mapped 2n-to-1 onto the “critical value ray” c p p v 0 c The critical points and prepoles all lie on the 1/2n “critical circle” r | | The critical circle is mapped 2n-to-1 onto the “critical value ray” And every other circle centered at the origin and outside the critical circle is mapped n-to-1 to an ellipse with foci at the critical values v 0 The critical points and prepoles all lie on the 1/2n “critical circle” r | | The critical circle is mapped 2n-to-1 onto the “critical value ray” And every other circle centered at the origin and outside the critical circle is mapped n-to-1 to an ellipse with foci at the critical values v 0 The critical points and prepoles all lie on the 1/2n “critical circle” r | | The critical circle is mapped 2n-to-1 onto the “critical value ray” And every other circle centered at the origin and outside the critical circle is mapped n-to-1 to an ellipse with foci at the critical values, and same inside v 0 There are no critical points outside the critical circle, so this region is mapped as n-to-1 covering onto the complement of the critical value ray. v 0 There are no critical points outside the critical circle, so this region is mapped as n-to-1 covering onto the complement of the critical value ray. The interior of the critical circle is also mapped n-to-1 onto the complement of the critical value ray v 0 The dividing circle contains all parameters for which the critical values lie on the critical circle, i.e., | |1/2n 2 | |1/2 | | 2 2n /(n1) The dividing circle contains all parameters for which the critical values lie on the critical circle, i.e., | |1/2n 2 | |1/2 | | 2 2n /(n1) QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n 4 : | | 2 8 / 3 When n = 4, the dividing circle passes through 3 centers of Sierpinski holes and 3 baby Mandelbrot sets The dividing circle passes through n-1 centers of Sierpinski holes and n-1 centers of baby Mandelbrot sets. QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n 4 : | | 2 8 / 3 When n = 4, the dividing circle passes through 3 centers of Sierpinski holes and 3 baby Mandelbrot sets Reason: v 8 / 3 | | 2 parameter plane n = 4 r | |8 dynamical plane Reason: as runs once around the dividing circle, v 8 / 3 | | 2 parameter plane n = 4 r | |8 dynamical plane Reason: as runs once around the dividing circle, v rotates 1/2 of a turn, v 8 / 3 | | 2 parameter plane n = 4 r | |8 dynamical plane Reason: as runs once around the dividing circle, v rotates 1/2 of a turn, while the critical points and prepoles each rotate on 1/8 of a turn. v 8 / 3 | | 2 parameter plane n = 4 r | |8 dynamical plane Reason: as runs once around the dividing circle, v rotates 1/2 of a turn, while the critical points and prepoles each rotate on 1/8 of a turn. So v meets 3 prepoles and critical points enroute. v 8 / 3 | | 2 parameter plane n = 4 r | |8 dynamical plane So the ring S0 is just the dividing circle in parameter plane. S0 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n=4 So the ring S0 is just the dividing circle in parameter plane. For the other rings, let’s consider for simplicity only the case where n = 4 S0 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n=4 When lies inside the dividing circle, we have | v | 2 | | | | |1/8 | c | When lies inside the dividing circle, we have | v | 2 | | | | |1/8 | c | so F maps the critical circle C0 strictly inside itself v v C0 Now there is a preimage C1 of the critical circle that is mapped 4-to-1 onto the critical circle, and this curve contains 32 pre-critical points and 32 pre-pre-poles. v v C0 C1 And then a preimage C2 of the C1 that is mapped 4-to-1 onto the C1, and so 16-to-1 onto C0, and this curve contains 128 pre-pre-critical points and 128 pre-pre-pre-poles, etc. v v C0 C1 C2 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. The rings C0 and C1 Let G( ) be the second iterate of the critical point G( ) F (c ) 2 2 n n/2 1 2n (n/2)1 Let G( ) be the second iterate of the critical point G( ) F (c ) 2 2 n n/2 1 So G( ) 16 16 2 1 2n (n/2)1 when n = 4. Let G( ) be the second iterate of the critical point G( ) F (c ) 2 2 n n/2 1 So G( ) 16 16 2 1 2n (n/2)1 when n = 4. Note that G() as 0 provided n > 2. 1 When n = 2, G( ) 4 , a very different situation. 4 G maps points in the parameter plane to points in the dynamical plane G QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. C0 the critical circle Let D be the open disk of radius 1/8 in the parameter plane. G maps D univalently onto a region in the exterior of C0 G(D) G QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. D C0 and G(D) covers C1, C2,... G QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. D C0 Choose a small disk D0 inside M. Then G maps the annulus A = D - D0 univalently over all of the Cj, j > 0. G A QuickTime™ and a TIFF (LZW) decompressor are needed to see this 0 picture. D C0 k Choose a parametrization of Ck, say ( ). So we k have a second map from A into G(A), ( ) G A QuickTime™ and a TIFF (LZW) decompressor are needed to see this 0 picture. D C0 k ( ) Since G is 1-to-1, we thus have a map H ( ) G 1 ( k ( )) which takes A into A. A QuickTime™ and a TIFF (LZW) decompressor are needed to see this 0 picture. H D C0 k ( ) Let S be the covering strip of A and let H*: S be the covering map of H: A A A QuickTime™ and a TIFF (LZW) decompressor are needed to see this 0 picture. S H D C0 k ( ) Let S be the covering strip of A and let H*: S be the covering map of H: A A S By the Schwarz Lemma, for each given k, , and , k there is a unique fixed point ( ) for H* in A, which depends analytically on . Let S be the covering strip of A and let H*: S be the covering map of H: A A S By the Schwarz Lemma, for each given k, , and , k there is a unique fixed point ( ) for H* in A, which depends analytically on . So the map () gives a parametrization of the ring Sk in the parameter plane, and -values that correspond to pre-poles or pre-critical points are then parameters at the centers of Sierpinski holes or baby Mandelbrot sets. k There are (n - 2)nk-1 pre-poles in the kth dynamical plane ring, but (n - 2)nk-1 + 1 centers of Sierpinski holes in the parameter plane rings. Here’s the reason: There are (n - 2)nk-1 pre-poles in the kth dynamical plane ring, but (n - 2)nk-1 + 1 centers of Sierpinski holes in the parameter plane rings. Here’s the reason: 1 1 On the annulus A, G( ) 16 16 16 2 There are (n - 2)nk-1 pre-poles in the kth dynamical plane ring, but (n - 2)nk-1 + 1 centers of Sierpinski holes in the parameter plane rings. Here’s the reason: 1 1 On the annulus A, G( ) 16 16 16 2 So as rotates clockwise around the ring Sk, G( ) rotates once around the origin in the counterclockwise direction.Meanwhile, each pre-pole and pre-critical point rotates clockwise by approximately 1/((n-2)nk-1 of a turn. There are (n - 2)nk-1 pre-poles in the kth dynamical plane ring, but (n - 2)nk-1 + 1 centers of Sierpinski holes in the parameter plane rings. Here’s the reason: 1 1 On the annulus A, G( ) 16 16 16 2 So as rotates clockwise around the ring Sk, G( ) rotates once around the origin in the counterclockwise direction.Meanwhile, each pre-pole and pre-critical point rotates clockwise by approximately 1/((n-2)nk-1 of a turn. So G( ) hits one additional prepole or pre-critical point while traveling around each Sk. Similar arguments show that each Sierpinski hole on a Mandelpinski necklace is also surrounded by infinitely many sub-Mandelpinski necklaces QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Similar arguments show that each Sierpinski hole on a Mandelpinski necklace is also surrounded by infinitely many sub-Mandelpinski necklaces QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Similar arguments show that each Sierpinski hole on a Mandelpinski necklace is also surrounded by infinitely many sub-Mandelpinski necklaces QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Similar arguments show that each Sierpinski hole on a Mandelpinski necklace is also surrounded by infinitely many sub-Mandelpinski necklaces QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Similar arguments show that each Sierpinski hole on a Mandelpinski necklace is also surrounded by infinitely many sub-Mandelpinski necklaces QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. Some open problems: Some open problems: 1. By Xiaoguang Wang’s result yesterday, the boundary of B is always a simple closed curve (except when J is a Cantor set) when n > 2. QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n=3 Some open problems: 1. By Xiaoguang Wang’s result yesterday, the boundary of B is always a simple closed curve (except when J is a Cantor set) when n > 2. Is the boundary of the parameter plane also a simple closed curve??? QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n=3 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n=3 Some open problems: 1. By Xiaoguang Wang’s result yesterday, the boundary of B is always a simple closed curve (except when J is a Cantor set) when n > 2. Is the boundary of the parameter plane also a simple closed curve??? QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n=4 QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n=4 Some open problems: 1. By Xiaoguang Wang’s result yesterday, the boundary of B is always a simple closed curve (except when J is a Cantor set) when n > 2. Is the boundary of the parameter plane also a simple closed curve??? 2. What about the crazy case n = 2??? Some open problems: 1. By Xiaoguang Wang’s result yesterday, the boundary of B is always a simple closed curve (except when J is a Cantor set) when n > 2. Is the boundary of the parameter plane also a simple closed curve??? 2. What about the crazy case n = 2??? 3. Are the Julia sets for these maps always locally connected? Some open problems: 1. By Xiaoguang Wang’s result yesterday, the boundary of B is always a simple closed curve (except when J is a Cantor set) when n > 2. Is the boundary of the parameter plane also a simple closed curve??? 2. What about the crazy case n = 2??? 3. Are the Julia sets for these maps always locally connected? 4. Are the parameter planes locally connected??? 5. What is going on in the parameter plane near 0 when n = 2? QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. 5. What is going on in the parameter plane near 0 when n = 2? QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. 5. What is going on in the parameter plane near 0 when n = 2? QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. 6. What is the structure in the parameter plane outside the dividing circle? QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. 7. What is going on in the parameter plane for the maps F (z) z n z QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n = 2, d = 1 Not a baby M-set 7. What is going on in the parameter plane for the maps F (z) z n z No Cantor necklace QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n = 2, d = 1 Not a baby M-set 7. What is going on in the parameter plane for the maps F (z) z n No Cantor necklace QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n = 2, d = 1 z 7. What is going on in the parameter plane for the maps F (z) z n J approaches the unit disk only along these 3 lines QuickTime™ and a TIFF (LZW) decompressor are needed to see this picture. n = 4, d = 1 z