BIOLOGIC RELATIONS BETWEEN PLANTS AND ANTS. 451 



early age, independently of the action of the ants. But it will always 

 be pertinent to ask why the ants did not cause the hypertrophy of 

 both the opposite leaves which ought a priori to be subject to the same 

 conditions as to the formation of vesicles. May it not be that primi- 

 tively without any intervention on the part of ants, there was a tend- 

 ency to an inequality of development between the two opposite leaves 

 of the same pair, as is seen in a number of vegetables that have oppo- 

 site leaves? The favored leaf would tend to form a vesicle which the 

 unfavored leaf could not do. This difference of size between the two 

 leaves of the same pair is still more marked in the Myrmedone macros- 

 permum of Brazil and the Galophysca heterophyUa of Peru. 



In Tococa truncata the foliary bursa is not much developed, but usu- 

 ally larger on one of the leaves than on its opposite. The inequality of 

 the same pair is very well marked in T.platypetala and T. bullifera. In 

 this last species the development of the limbus seems to be propor- 

 tional to that of the bursa. In the section Epiphysca of the genus 

 Tococa the bursa is evidently formed at the base of the limbus. In 

 the section Hypnophysca it seems to be produced from the petiole, but 

 in reality it is the limbus that is decurrent along the petiole and forms 

 there an elongated bursa. The fact is clearly shown in Tococa bullifera 

 (fig. 10). The bursae of Tococa for micaria, T. guianensis and T.platy- 

 petala are very fine. The stem of T. guianensis and formicaria seem to 

 be hollow. 



The relation between the size of the limbus and that of the bursa 

 leads us to suspect that the internal surface of the bursa may be 

 endowed with absorbent properties. If the bursa played the part of a 

 true absorbing organ it would yield to the limbus nutritive material, 

 which would explain why a leaf possessing a larger bursa also posesses 

 a larger limbus. In dried specimens we find in the bursa a large 

 quantity of detritus. Tbe internal surface of T. formicaria and T. 

 guianensis bristles with papillae and hairs. In the first species we find 

 there scale-insects as well as ants. 



Beccari found entire colonies of ants with pupae on single specimens 

 of T. bullifera and Myrmedone macrosperma of Brazil and Venezuela. 

 These plants have bursae at the base of the limbus, -which appear more 

 complete than those of other species. The transverse nervures that 

 run over them project on the inner surface as lamellae that indirectly 

 divide the bursa into galleries, as in the tubercle of the Hydnopliy turns. 



In Majeta guianensis the internal wall of the bursa is lined with elon- 

 gated papillae formed of cells normally filled with a colored protoplasm 

 that seems analogous to the foliary glands of Drosera. This fact, 

 together with the presence of fragments of ants and other insects, has 

 led Beccari to suppose that the bursa may in this case have digestive 

 and assimilative functions. 



Some species of Tococa have leaves destitute of bursae. In others 

 (T. subnuda) the bursas are rudimentary. The examination of this 



