12G MKKISTIC VARIATION. [PARTI. 



of several cases already given, it is not impossible that it' the 

 variation -< -n in this frog had gone further, the 10th vertebra 

 iiii^lit alone support the ilium (cp. Nos. 57 and GO) and thus 

 pre-.-nt the characters of the normal 9th in their completeness. Jf 

 thi- change had taken place, we should have a case like that of 

 1,'i'i'ilin, and there would be nothing to shew that the new loth 

 l.ra was not the 9th of the normal. The truth then seems to 

 I- that owing to the correlation between Meristic Variation pro- 

 ducing change in number, and simultaneous Substantive Variation 

 producing a change of form or rather a redistribution of characters, 

 tin- attempt to trace individual homologies must necessarily fail ; 

 t-r while such determination must be based either on ordinal 

 PMtion or on structural differentiation, neither of these criterions 

 are really sound. As I have tried to shew, the belief that they are 

 SO depends rather on preconception than on the facts of Variation. 



;,7 A male -pe.'iineii of R. tti/i/</-<i/-i g with tnt free vertebra- is 



d. -eril.ed t,\ Ho\\i:s. In this case the 9th had a posterior 



- "ii the left side only. Upon the left side the 



isveree process of the 9th was not larger than that of the 8th 



and did not >upport tin- ilium, which on the left side was entin-lv 

 1-orii'- 1-y tip- large transverse proct-ss of the 10th. On the right 

 -id<- the tran.-\er-e processes of both 9th and 10th were developed 



to -uiiport th'- ilium, neither being in it-elf so large as that of the 

 loth on tip- left side. Tin- !Mh \\a- concave in front instead of 

 convex a- ii-ual. and thus tin- Mh which i> normally biconcave is 

 ci.nv-x In-hind. Tin- posterior t'aci-s of lotli !Uh and 10th bore two 

 convexities Midi as .-in- normal to the !th. The nrostvle was 

 normal, having \\.-ll-dev.-loped aperture^ for exit uf the last pair 



of s|iinal Derves. Il<i\vi.s. <;. i;., Aunt. Az., i. ISSG, p. ^77, 



ii< I 



In thi- case th.- d.-partur- from the normal, exemplified by 

 No. .".;, lia> gone -till furtln-r, and the new 10th vertebra bears the 

 ilium wholly ,>\\ th<- 1- tt *[<{< and in part on the right. The con- 

 dition U thu- again int.-rmcdiat.- between the normal and a com- 

 plct.- traii-torination ot tin- !th into a trunk vertebra and the 

 introduction .,(' a 10th to b.-ar the ilium (as in Xo. GO). As 

 regard > the hoinologi.-s of the vertebrae, the same issues are 

 a-ain raided which w.-n- indicated in regard to No. 56. 



Rana teraporaria : Case in \\hi.-h transverse processes were present 

 in tin- atlas \.-rtcln-a and the transverse processes of several of the 

 vertebra were al. normal (Fig. 11. [.). The atlas possessed well- 

 developed t ran ^\ . !> pr. .^-i-v. 



In i lie axis the transverse processes are direete.l forwards instead 

 Of Kackwank and that of the I. ft -Me presents an indication of bifur- 

 tion al M - e\t remit v. 



'I'll.- third \.-rtel, ra j .,1 two pair- ,.f transverse processes 



which are joined together for two-thirds of their length. The fourth 



