1907.] Wheeler, Fungus-growing Ants of North America, 771 



removing the stone or piece of wood the galleries are seen to be very irregular, 

 running along the surface as in the nests of C. icheeleri and extending down 

 into the soil to a depth of 20 to 35 cm. They adapt their course to the many 

 small fragments of limestone on or below the surface. The single fungus 

 garden, of irregularly flattened or sometimes of elongate and straggling 

 form lies in dilated portions of the gallery, usually completely exposed by 

 the removal of the stone. In many nests the garden rests on a small stone, 

 piece of bark or dead leaf from which the earth has been carefully removed 

 by the ants. So different is this garden from that of the other Attii hereto- 

 fore described that it has been completely overlooked by all previous ob- 

 servers. The substratum consists of a mass of caterpillar droppings a few 

 cm. in diameter, Avhich have undergone so little manipulation by the ants 

 that the individual pellets may be distinctly recognized even to the pecu- 

 liar ridges produced by the rectal folds of the caterpillars. 



The fungus grown on this substratum is not a mycelium as in all the 

 species above described, but is in the form of a number of isolated whitish 

 or yellowish bodies .25-.5o mm. in diameter, of the appearance and consist- 

 ency of cheese crumbs and of an irregularly polygonal or p\Tiform shape 

 (PI. L, Fig. 29). Each of these bodies may be said to correspond to a cluster 

 of gongylidia and may therefore be called a bromatium. It rests with one 

 of its angles or surfaces on the caterpillar excrement, but no rhizoids or 

 mycelial threads can be seen at this point entering and ramifying in the 

 substratum. The whole garden is kept so moist that when first exposed to 

 the air the surface glistens with a film of greenish licjuid. As the bromatia 

 rest on this liquid, which evidently represents a thick solution of fecal and 

 vegetable substances, they are in a position to absorb nutriment directly. 

 It is probable that the habit of placing the excrement on the surface of a small 

 stone, bit of wood or dead leaf which happens to be found in the gallery of 

 the nest, is for the purpose of retaining this nutrient moisture and pre- 

 venting its absorption by the soil. All of these conditions are such as to 

 restrict C. comalensis and the other forms of rimosus to moist, shady localities. 

 Such situations are of course, also the only ones in which tropical and sub- 

 tropical plants are suflSciently abundant to furnish an unfailing supply of 

 caterpillar droppings. 



When the bromatia are crushed and examined in water under a high 

 power of the microscope, they are seen to consist of a dense mass of elliptical 

 or subspherical cells measuring .78-2 jx in length and .78-1 pt in breadth. 

 Among these there are also cells of other shapes and even smaller sizes as 

 shown in PI. LI I, Fig. 43. The c}1oplasm of all of these cells is colorless 

 and finely granular and contains one or more clear vacuoles and a few small 

 refractive corpuscles. A nucleus is probably present, but I have been 



