VARIATION AND ZOOGEOGRAPHY OF LIGUUS IN FLORIDA. 437 



most or all of the yellow color, but with renewed growth the yellow appears along 

 the growing edge, as shown in PI. XXX\ 1 1 1 , fig. 1 8, et c. Youngshells of xant 1 stic 

 races are therefore more strongly colored than the corresponding whorls of any 

 of the adult shells of the same colonies. This fading of the yellow pigment in 

 living shells is the more remarkable because in dry museum specimens collected 

 over twenty years ago, the yellow still remains. In Say's type of Liguus solidus 

 the yellow tint may still be seen, although it was collected in 1824. ^ How . a 

 new feature in I loridan Liguus. There is rarely any trace of it in Cuban forms. 

 Yellow and white grounds are probably a pair of Mendelian characters in Floridw 

 Liguus. In inheritance yellow is evidently independent of the rest of the color- 

 pattern. 



The chestnut-brown and blackish pigment is in the prismatic layer of th< 

 shell, and is the most permanent, traceable in dead, weathered shells aft< r other 

 colors arc lost. Brown intensified becomes almost black; overlaid with a film 

 of the white ground-color it becomes purplish or bluish. It is ne\< int< * 

 changeable with yellow, in Liguus, however the case may stand in vert eh ratal 

 This is obvious in most of the colonies. Thus at Key Vaca (PI. XX? VII, fU 

 9-9d) the brown pattern varies independently of yellow, which ma; be int< 

 weak or absent in forms with maximum or minimum brown pattern or none at 

 all. While the chestnut zones overlay yellow zones in most cases, yet in 

 Liguus the darker color cannot be considered to be merely a further stage 

 of oxidation of the yellow pigment. The two are absolutely distinct in inherit- 

 ance and phylogeny. The -patterns of black or brown never appear in ellow. 

 Whether the inheritance factor for dark color is or is not separable from that for 

 flame pattern under one guise or another, it is certain that it is never diwociated 

 from it in Liguus and related genera, and equally certain that ycllov nc\ r is 

 controlled by that pattern. 



Rose or pink coloring of the early whorls, associated with ros pink or purple 

 on the parietal wall and columella, are highly characteristic of Liguus fasciatus 

 and seem absolutely constant in pure races of that species and its derivative 

 varieties. Rose is apparently dominant in hybrids. 



The primitive color-pattern of Liguus must have consisted of brown longitud- 

 inal stripes or flames on a white ground. This conclusion rests upon the fact 

 that this pattern characterizes most of the species of all the genera of the sub- 

 family Orthalicince to which Liguus belongs, at least in their young stages 

 It is highly improbable that the same fundamental pattern would have been 

 independently acquired by species of a half dozen related genera, 7 but far mor<- 

 likely that they inherited it from a common ancestral stock. If this be so, thei 

 the spiral band patterns are secondary. This does not apply to the green lines, 

 which have no connection with the flame pattern. 



The dark flames usually appear on the last embr; >nic whorl as continuous 



7 Manual of Conchology, XII 

 in other sub-families of BulimtU\ 



Similar flame or streak 



