464 ANNUAL, REPORT SMITHSONIAN INSTITUTION, 1910. 



the sea. Yumiiri Valley would constitute just such a harbor were it 

 submerged beneath the sea. 



This study of the reefs of Cuba and the Bahamas naturally led 

 him to renew his observations in Florida and to visit the Bermudas. 

 He saw the Bennuda Islands in March, 1894, and in December of the 

 same year he chartered a tug and steamed along the Barrier Reef of 

 Florida. 



He found that, in common with the Bahamas, the Bermudas con- 

 sist of a?olian limestone. In places the interior of these islands as 

 dissolved away by the action of rain water rendered acid by decom- 

 posing vegetable matter, and thus depressions were formed in the 

 central parts of the islands. Then when the islands sank the sea 

 broke through the rims and filled the lagoons, afterwards deepening 

 them by its scouring action. 



Thus the Bermudas have assumed an atoll-like shape, but their 

 contour is not due to corals. Indeed, there are but feAv corals at 

 Bermuda, and these form a mere veneer over the sunken seolian 

 ledges. The so-called miniature atolls are mere pot-hole basins 

 which have been scooped out by wave action in the seolian rock, and 

 their rims are never more than 18 inches high, and consist of a wall 

 of a3olian rock covered by a coating of serpulse, algse, and corallines, 

 which enable them to withstand the wearing action of the sea. Thus 

 Darwin's theory of coral reefs can not explain the conditions seen 

 in the Bahamas and Bermudas. 



The results of his study of the Florida Reef were finally pub- 

 lished in 1896, in cooperation with Dr. Leon S. Griswold. Agassiz 

 concludes that the Marquesas of Florida are not an atoll, but inclose 

 a sound that has not been formed by subsidence, but by the solvent 

 and mechanical action of the sea. Thus the Marquesas are similar 

 in their geological historj^^ to other sounds back of the line of the 

 Florida Keys. 



He found an elevated reef extending along the seaward face of 

 the Florida Keys from Lower Matacumbe to Soldier's Key. We 

 now know, however, that the elevated reef actually extends from the 

 southern end of Big Pine Key to Soldier's Key. Agassiz believed 

 that the oolite limestone back of the elevated reef and along the 

 mainland shore of Key Biscayne Bay was seolian rock ; but Griswold 

 decided that it was only a mud flat which had been formed beneath 

 the water, and afterwards elevated. Later studies have shown that 

 Griswold was right. 



In 1895 he instituted a study of the underground temperature of 

 the rock walls of the Calumet and Hecla mine, and found that the 

 increase is only 1° F. for eveiy 223.7 feet as we descend. His deep- 

 est temperature observation was 4,580 feet beneath the surface of the 

 ground. 



