VARIATION AND ZOOGEOGRAPHY OF LIGUUS IN FLORIDA. 433 



crenatus for example can be directly traceable to the action of physical environ- 

 ment. 



Fluctuating variations among individuals of a pure colony are not great. 



Such conspicuous "mutations" as I have noticed in colonies of pure races belong 

 to the class of varieties produced by loss of a character. Thus in L. /. lignumvitir, 

 PI. XXXVII, figs. 4, 4a, 4b, the mutation 4c, 4d has been formed by loss of the 

 dark color-pattern, the roseate apex and green lines remaining unchanged* 



L. solidus has produced similar mutations by loss of some factor controlling 

 the deposition of dark color- pattern. Compare PI. XXXVII, figs. 1, la with \h. 

 A few similar cases have been noticed in other pure coloni* While the 



parentage of these forms is of course unknown, their sporadic oocurrenee unac- 

 companied by intergrading individuals in otherwise pure colonial may warrant 

 the inference that we have to do with mutations w h ( re by loss of a "color-factor" 

 a considerable change has been effected abruptly. .Under natural conditions now 

 forms so produced will result in races only when ganietically pure individuals 

 happen to be isolated from the parent race — an unlikely but possible contingency. 

 I do not think that natural selection or sexual selection enter into the question 



in these snails. 



There is evidence that pattern undergoes gradual change in succetiive 

 generations, beginning with the later whorls, working backward, affecting the 

 early post-embryonic whorls last. Thus in castaneozonatus there are colonies 

 in which the dark pattern extends upon the last whorl (PI. XXXIX, fip 

 22a-d, 23, 23a), others in which it is restricted to the spire (PI. *X\IX fifes. 

 20g-j). In decadent series the last remnants of the pattern always remain 

 on the early whorls, never on the later (PL XXXIX, fig. 20/'), as if ito loss 

 were due to decreasing potency of the factor or factors for brown color ae 

 the individual grows older. I do not doubt that new color-races have been 

 produced by this process; and as its action seems to affect a whole colony, 

 the resulting races would have the advantage (in species-forming) of gametic 

 purity, which would not be possessed by similar forms produced by the 

 mutation process. Probably castaneozonatus has been converted into r uatuM 

 in this way. In progressive series the early whorls are the last to be affected by 

 changes initiated on later whorls. Thus in castaneozonatus the complete coales- 

 cence of brown flames into continuous zones is first manifest on the later whorls 

 (PL XXXIX, fig. 22a), but only in the most progressive (or as Hyatt tern it, 

 accelerated) colony has this coalescence invaded the early whorls (PL XXXI \, 

 figs. 23, 23a, Key Largo). Hyatt, Branco, Grabau and others have traced act ual 

 phylogenies of cephalopods and gastropods in which the changes proeeed m this 

 fashion. Much therefore is to be said in favor of evolution by variation in 

 the potency of the character-determining factors during ontogeny, combined 



4c. 4d would technically be L. /. rotealu* so far as color is ronatrned but various 



peculiarities in the remaining pattern, as well as the shape and light texture of the shells dfeow that 

 are dealing with a mutation of the L. /. lignumvit*, in a colony of which thete were found, no other 



of Liguus inhabiting the sam 



JOURN. ACAD. NAT. SCI. PHILA 



