132 PROCEEDINGS OF THE ACADEMY OF [1889. 



in Cumberland. 1 Specimens brought to Philadelphia, however, and 

 carefully examined chemically and microscopically go to prove that 

 these tubes are not the result of lightning strokes, and that the 

 hardened interior is not vitrification, but solidification due to 

 partial solution of the loose grains of sand. 



Evidence is thus obtained to show that these tubes represent the 

 ramifications of plant roots of a now exterminated flora ; plants 

 which existed probably upon the first sands drifted upon the original 

 rocky core. The juices of the roots acting upon the sand inmediately 

 surrounding them formed a compact layer. Through erosion and 

 subsidence the vegetation was afterwards exterminated, the looser 

 particles of the drift rock worn away and the surface left covered by 

 myriads of tubes of all sizes formerly occupied by plant roots and 

 rootlets. The small islands exhibiting these peculiar formations 

 are indications, therefore, of erosion and subsidence, and in the 

 Bahamas, the processes of land destruction and land formation may 

 be seen actively at Avork. There is reason to believe that the 

 topography of the present islands has undergone great change, and 

 the probabilities of finding a locality in which the present flora 

 represents an undisturbed indigenous growth are slight. 



Fossil corals of recent types are found much above high water- 

 mark ; large caves with their floors above tide level exist, together 

 with many other evidences of land elevation. Many of the islands 

 show the formation of seolian rocks in progress, while in the great 

 bights dividing Andros Island, and elsewhere, the deposition of for- 

 aminifera has been so great that former channels and sponging 

 grounds have within very recent years become too shallow for ap- 

 proach. On the whole the Bahamas seem to be growing, but changes 

 in the configuration of the group have also undoubtedly been going 

 on through erosion and local subsidences probably owing to the 

 undermining action of the waves.' 2 



The role played by vegetation in determining the character of 

 land surface is again well shown in the so-called " Banana holes," 

 so abundant in New Providence and other islands, holes varying in 

 size from that of a pint cup to that of a large cistern. They are 

 suggestive of pot holes, but can have no such origin and are evidently 

 not cut out by the waves at any previous period of subsidence. He 

 could account for their formation in but one way, and that is 

 through the action of decaying vegetable matter. Each of these 

 holes contains large quantities of leaves and other vegetable sub- 

 stances which, being kept wet by the heavy rains and by the fresh 

 water elevated by each rising tide (almost all wells have a regular 

 ebb and flow in these porous islands), undergo fermentative changes, 

 by the products of which the soft calcareous rock is dissolved and 

 leaches away. He was interested also in examining sections of ex- 



1 Geol. Trans., Vol. ii, p. 528. 



2 Vid. Fewkes, on the Origin of the Present Form of the Bermudas, Proc. 

 Boston Soc. Nat. Hist., Vol. xxiv, 1888, p. 518. 



