152 HYMENOPTERA. 



crude silk, and the threads could be seen crossing each other, the 

 same structure being observed at the top and bottom of each 

 cell. 



In the three-celled nest of Polistes (Plate 5, Fig. 5, 5 a) 

 first noticed April 29th, there were but two eggs deposited, the 

 third cell being without an egg, and a little smaller, and 

 the rim not so high as in the other two. The outer edge did 

 not seem to be perfectly circular, though stated by Water- 

 house to be so in the incipient cells, for in some cases we de- 

 tected two slight angles, thus making three sides, which, 

 however, would be easily overlooked on casual observation ; 

 as there ai*e only two sides within, the cell, from being at its 

 earliest inception hemispherical, or "saucer-shaped," becomes 

 five, and subsequently six-sided, and thus from being cir- 

 cular, it is converted by the wasps into a hexagonal cell. In 

 some cells, perhaps a majority, both in this and the other spe- 

 cies, the newl}^ made rim of the small cells is thinner than the 

 parts below, and slightly bent inwards ; thus being quite the re- 

 verse of the thickened rim of the cells of the Hive Bee. It 

 would seem that the wasp plasters on more silk, especially on 

 the angles, building them out, and making them more promi- 

 nent, in order to complete, when other cells are added, their 

 hexagonal form. The three cells are of much the same size 

 and height when the third egg is laid, as we observed in another 

 nest, that of Polistes Canadensis (Linn.), built at the Defences 

 of Washington, near Munson's Hill, June 9th. 



Again, when one or two more cells have been added to the 

 nest, and there are four or five in all (Plate 5, Fig. 6 ; 6«, top 

 view, in which there are four cells), two of them are nearly 

 twice as large as the others, while the fifth has been just begun, 

 and is eggless. The form of the two which run up much higher 

 than the others is the same as that of the smaller and shorter 

 ones, i. e. they are on one side nearly semicircular, and on the 

 other, parti}' hexagonal, and the angular sides show a tendency 

 to be even more circular than when the others are built around 

 them, for the little architect seems to bring out the angles 

 more prominently when carrying up the walls of tlie other cells. 

 Thus she builds, as if by design, one and the same cell both 

 by the "circular" and "hexagonal" methods, afterwards adopt- 



