320 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1911. 
In view of the habits of ants, it seems highly probable that at the 
time of leaving their nests the young females scatter over the surround- 
ing region pretty much at random. When they alight, some of them 
find themselves in locations where ant colonies, on account of the 
character of the ground, can not possibly survive, and as these young 
females break off their wings as soon as they alight, they can not 
renew their flight and seek more favorable ground, but they must 
perish without having founded a new colony. And this must happen 
over and over again, with the final result that localities unfavorable 
for ants do not have ant colonies, while the favorable localities may 
have a superabundance of them. Favorable and unfavorable con- 
ditions are not always sharply defined, but merge into each other. 
In some cases it is quite evident what constitute unfavorable condi- 
tions. Ground that is constantly wet or liable to inundation can not 
be occupied; hard, rocky surfaces, or even very thin soils, are not 
available; soils so sandy or friable that underground tunnels dug in 
them will not stand are evidently not available for the establishment 
of colonies. 
Between soils most favorable and unfavorable ones there are all 
sorts of gradations, so that one is prepared, for this reason alone, to 
find the ant hills bigger and more abundant in some places than in 
others. It is evident that it is all a question of adaptability, how- 
ever, rather than a matter of choice on the part of the ants. 
Just what kind of soil is most favorable for the ants I can not state 
positively. My general impression is that the mounds are most 
abundant on clayey soils, whether the clay comes directly from the 
decomposition of feldspathic rocks or from the disintegration of shales 
and slates. | 
This preference for the clayey soils is well shown at many places 
through the diamond-bearing highlands of<the interior of Bahia, 
where the diamond-bearing quartzites, known as the Lavras series, 
are underlain by a thick series of shales called the Caboclo series. 
The Lavras beds heing quartzites, or sandstones, break down into a 
very sandy soil, while the Caboclo shales form a stiff, clayey soil, and 
as they are adjacent to each other the line of demarcation between 
the two soils is usually well defined. While traveling through that 
district in 1907, I was frequently able to locate myself geologically 
by the abundance or absence of the ant hills. Not infrequently the 
line of parting between the two series was concealed by a thick soil 
and overgrown with forests, but the distribution of the mounds 
would often show the line of parting within 20 or 25 meters. 
My former assistant, Mr. Roderic Crandall, who has traveled exten- 
sively in Bahia, Pernambuco, Piauhy, Minas, and Goyaz, writes, in 
reply to my inquiries, as follows regarding the preference of the ants 
for certain soils: “In Bahia the ants of all kinds show a preference for 
