VARIATION AND ZOOGEOGRAPHY OF LIGUUS IN FLORIDA. 419 



now predominate over those making land whether by elevation or otherwi* 

 On the lower keys which I have explored, all stages in the process of disinte- 

 gration are readily observable — flask-shaped "cisterns" of fresh rater in th< 



consolidated rock, sinks produced by their caving in, finally channels from 



caving of the subterranean water courses, etc. 10 Professor \lcxander Agassis, 10 



who examined the eastern keys, writes: "It is most probable that Key Biscaym 

 Bay, Card's Sound, Barnes Sound and the smaller sounds to the north of Key 

 Largo, as well as a number of ill-defined sounds between Barnes Sound and the 

 B: of Florida, owe their origin to the erosive and solvent set ion of the sea. 

 A glance at the chart clearly indicates the former connection with the mainland 

 of the line of keys extending from Key Biscayne to Long Kc\ " The tongu 

 of land connecting Key Largo with the mainland, tin Rubicon and A niel%<f 

 Keys, and the shoals inwards from the Ragged Keys are mentioned h\ Agassis 

 as remnants of the former land. The "sounds may have originally been sinks 

 [formed by the action of rain water] similar to those of the Bermudas and Ba- 

 hamas, and have, like those of these islands, been changed into sounds by th< 

 breaking through of the barriers separating them from the open sea. This 

 would allow the formation of more or less deep channels . . . through w ieh 

 material held in suspension or solution could pass out. 



We have to do therefore with a region which has undergone considerable 

 changes in post-Pliocene time, ending wit h a period of dis hit ion st ill in progress 

 during which the keys have become progressively more fully isolat 1 from the 

 mainland and from one another. This progressive disintegration of the land 

 areas, which seems established beyond question, enables us to fix the relative 

 dates of the successive migrations of Liguvs with a high degree of probability. 



1. Liguus crenatus (PI. XXXVII, figs. 5-86) occurs as a pure race only in 



the most isolated localities (fig. 4, page 442). At Ix>ssman'a Key on th- 



>» 



western border of the Ten Thousand Islands. On New River (fig. 3 I where it 

 is isolated by a long st retch of pine and palmetto land. On Elliott 's Key . isolat* 1 

 by the deep Caesar's Creek, and on the keys south of Largo. A glance at the 

 map (PL XL, and fig. 3, page 441) will show that these are peripheral areas, which 

 must have been among the first to be isolated. It appears therefore that 

 crenatus is an older race in its region than the more highly colored races, which, 

 coming later, found the paths to the peripheral keys barred by channels. Prob- 

 ably crenatus overran the area made available by the elevation of the r< fs, as 

 soon after that event as suitable woodlands became establish 1. The lai 1 

 surface in southern Florida must then have been far grea* r than at present. 

 The subsequent isolation of the scattered colonies of pure L. crenatus has permit ted 

 the formation of local races, as noted in the descriptions of the colonies. On the 

 mainland, where later migration was unrestricted, the crenatus territory has 

 been invaded by other races, forming hybrid colonies. 



• For observations on the formation of sinks and lakes by solution, with estimate of material 

 removed, see E. H. Sellards, Third Annual Rep. Florida State Ged. Survey, p. 60. Alio MS A. 

 Agassis, The Florida Elevated Reef, BuU. Mus. Comp. ZooL, XXVIII, No. 2, 1896. 



