408 BUREAU OF AMERICAN ETHNOLOGY [Bull. 187 



Ribault were thatched with reeds, by which he probably means cane 

 mats, while those along the St. Johns and to the south and west were 

 covered with palmetto. The change also seems to have been accom- 

 panied with a change in the interior structure of the roof, the roofs 

 to the north being of mud and wattle like the walls, while those to 

 the south were more open to the air for reasons connected with the 

 latitude. We read in Biedma's narrative of the De Soto expedition, 

 "there was a change in the habitations which were now in the earth, 

 like caves: heretofore they were covered with palm leaves and with 

 grass." (Swanton, 1922, p. 353; Bourne, 1904, vol. 2, pp. 9-10.) 

 This was near the lower course of Flint River. The Fidalgo of 

 Elvas places this change at approximately the same place, the houses 

 in the province of Toalli (later Tamathli) being the first to show 

 it. As the roofs are said to have been of "canes in the manner of 

 tile," we seem to be confirmed in the assumption that the cane was 

 woven into mats. Before that he says that the houses had been 

 roofed with "hay," by which he may mean palmetto, though it is 

 possible that grass was used in an intermediate area between the 

 palmetto country and the region where mats prevailed (Robertson, 

 1933, p. 74). 



The only strictly summer houses alluded to by early writers are 

 rude arbors into which we are told the Timucua retired during 

 3 months of the year. They perhaps furnished the original sug- 

 gestion for the open-sided Seminole houses described elsewhere. 



There is no account of a Calusa house, but Dickenson has left 

 us a brief sketch of the house of the chief of Santa Lucia on the 

 southeast coast. This was rectangular like that of Timucua chiefs, 

 but the main entrance was midway of one of the longer sides instead 

 of at an end. It bears a rather striking and surprising resemblance 

 to the Cusabo chief's house, and this might be considered of some 

 historical significance if the details were not so meager. (Dicken- 

 son, 1803, p. 33 ; Swanton, 1922, pp. 64, 391 ; S. C. Hist, and Genealog. 

 Mag., 1904, vol. 5, pp. 57-82.) At some earlier period it is believed 

 that pile dwellings were in use on the southwest Florida coast, but 

 nothing is known regarding them, except what may be inferred 

 from the findings of Gushing at Key Marco : 



None of the piles found by us exceeded six and a half feet in length. Indeed, 

 the greater number of them were less than three and a half feet long. These 

 shorter piles were nearly always made of palmetto wood, were not round, but 

 broad, or somewhat flattened, although the edges were rounded. They were 

 tapered toward the bottom and bluntly pointed, rudely squared or hollowed out 

 at the tops as though to support round, horizontal timbers ; and they were bored 

 or notched slantingly here and there through the edges, as though for the recep- 

 tion of rounded braces or cross-stays of poles or saplings, abundant pieces of 

 which were found. Some of the piles were worn at the points or lower ends, 



