THE FOUNDING OF COLONIES BY ATTA SEXDENS. 361 



easily seen. It may be said that the devouring of eggs is a very 

 common occurrence. I have seen it take place as often as six times 

 in two hours — in fact, as often as four times in a single hour — and 

 have noticed it in the case of all the Atta females that have been 

 under my observations. According to my investigations, it may be 

 stated that during the early breeding period an Atta female will, on 

 the average, lay not less than 2 eggs every hour — that is, approxi- 

 mately 50 eggs a day. But as has already been stated, during the 

 first ten or twelve days the number of eggs increases onh^ about 10 

 daily, so that for each 5 eggs laid 4 must be eaten. If the total 

 of the eggs for the period of brood development is estimated up to 

 the time of the appearance of the first working ants — let us say a 

 minimum of forty days — we get a total of 2,000 eggs, whereas the 

 entire brood at this time, including unhatched eggs, larvae, and 

 nymphs, will not exceed 200. "We therefore find a proportion of 

 eggs devoured to every 10 laid. 



The foregoing especially unfavorable proportion is doubtless to bb 

 attributed to the fact that the larva? required from the beginning 

 to be fed with the eggs. The feeding of larvae is a process somewhat 

 more difficult to observe than that of manuring the fungus-garden, 

 the ^gg laying, or the devouring of the eggs on the part of the 

 female, because it only rarely happens that the larvae are in a posi- 

 tion to be accurately observed; but on several occasions I have been 

 fortunate enough to have this procedure, from beginning to end, 

 under inspection by means of a hand lens. When the Qgg has been 

 laid, the mother-ant tests it for some seconds by a process of tasting, 

 and then turns to one of the larva*, which she caresses with her 

 antennte until it begins to move its mandibles. The Qgg is then thrust 

 between its mouth parts with considerable force. These continue to 

 work back and forth upon the Qgg^ which either stands perpendicular 

 to the larva, or, as is more frequent, lies along its ventral side. In 

 the latter case the mother-ant often presses the egg Avith her foot 

 against the larva. If it is still very young, the egg is generally after 

 a time taken away and given to another larva. A good-sized larva 

 is, however, capable of devouring an egg in from three to five min- 

 utes, so that nothing but the leathery skin remains, Avhich is later 

 removed by the mother-ant. At an}' rate, I have clearly observed 

 that a larva whose mouth parts were in vigorous action upon an 

 empty egg-skin had this residue licked away bj^ the mother-insect, 

 and that then the movements of the larva ceased. It is no doubt 

 due to the rapidity with which the larvae devour the eggs that one 

 so very rarely comes upon them in the actual process of eating them, 

 but I have clearly established the fact that feeding of larvie with 

 eggs is a very frequent occurrence. Thus, for example, I have noted 

 in a forenoon the process of Qgg laying to take place four times and 



