48 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. 
grove roots and stems are still in place in the molian marl near the land- 
ing to the quarry. There we find a thick bed of marl, — very similar to 
that off the west shore of Andros, — but it is filled in all directions with 
stems and roots and branches of mangroves, that form when decom- 
posed the tubes mentioned above. It appears as if this marl had been 
formed by the blowing of masses of coral sand into the mangrove bushes 
growing in shallow water; the sand soon became consolidated, or in- 
differently so, and in this mass of marl the mangroves either con- 
tinued to grow, or became overwhelmed by the small dunes. The marl 
off Andros may possibly have been formed in this way. Here the marl 
is full of small shells such as live in shallows, where they have been 
overwhelmed by the sand blown over them from the reef beaches. Тһе 
marl off the south beach of Boca Chica, and at our anchorage in 
twelve feet of water, is similar to that of the west shore ; it is probably 
swept out by the tides, and is also carried northward by them. 
Тһе condition of the northern part of Boca Chica is interesting, as 
showing what may probably have taken place all along the southern 
face of Florida, wherever there were sinks separating the flats of the 
elevated reefs, or deep patches or pools into which the beach sand de- 
rived from the reef could be blown. This solian sand thus formed in 
shallow water deposits like marl, which gradually filled the mangrove 
swamps, and when rising above the surface formed low жоПап hills 
standing in the very midst of reef patches. 
With the light thrown upon the mode of formation of sounds along 
the northern portion of the Florida Keys, as well as from the formation 
of secondary sounds by the erosion and disintegration of smaller islands, 
like those to the eastward and westward of Key West as far as the 
Pine Keys on the one side and Boca Grande on the other, an entirely 
different conception of the formation of the Marquesas has been sug- 
gested. 
I am now inclined to look upon the Marquesas as merely a broad 
circular sound, more complete than any other sound formed on the 
line of the keys, and not as an atoll. Being isolated, it is more strik- 
ing than other similar circular sounds, which are more or less sur- 
rounded by incomplete sounds, spits, or isolated parts of islands, as in 
the Pine Islands and the islands immediately to the eastward of Key 
West (Plate XV.). 
The islands forming the Marquesas group (Plates TX., X.) are sepa- 
rated by well marked passages, which I had formerly looked upon 88 
the outlets of an ordinary lagoon till they had been raised well above 
