I have often been asked whether I could give an approximate estimate of the quantity of seeds contained in a nest of average size, but I have hitherto felt unable to do this in a satisfactory manner. I am now in possession of more reliable data, and believe that the following calculation may be taken as a near approximation to the truth. During the spring of 1873 I removed with but very little loss the contents of two granaries from a very extensive nest of _Atta structor_, consisting principally of seeds of clover, fumitory, and pellitory. These seeds, when perfectly clean and freed from earth, weighed in the one case 4 sc. 4 grs., and in the other 5 sc. 8 grs. Now there cannot have been less than eighty such granaries in this nest, so that, if we take five scruples as the average weight of the seeds in each granary, and this, allowing for loss in collection, which we may fairly do, we should have a total weight of more than sixteen ounces, or one pound avoirdupois weight of seeds contained in the nest. But, though this mass of seeds represents the result of infinite labour on the part of the ants, each individual granary contains but an insignificant quantity, and the store-chambers often lie at great distances apart; it is therefore impossible to believe that the stores alluded to in the Misna can have been as small and scattered as these were, and we must, on the contrary, suppose them to have been both larger and more accessible.[115] [Footnote 115: Perhaps these heaps of corn may have been piled up at the entrance to the nest, as is sometimes the case when the workers, in their eagerness to secure as much as possible of a passing harvest, bring in the supplies too fast for their companions within the nest to be able to find room for and accommodate. When this happens the seeds lie outside the nest until fresh chambers are prepared for their reception.] The means employed by the ants to prevent the germination of the seeds contained in their granaries still remain secret, and all the experiments and investigations which I have hitherto been able to make have failed to give me the clue. The problem to be solved is the following: Given seeds, the readiness of which to germinate has been proved, to place them in damp soil at depths varying from half an inch to twenty inches below the surface in such a manner that they shall remain there dormant, neither germinating nor decaying, for weeks and even months. These very seeds must be capable of germinating after the conclusion of the experiment.