490 INSTINCT OF INSECTS. 
range where the male cells commence: and in the same 
manner, when they wish to revert to the modelling of 
the cells of workers, they pass by a gradually decreasing 
gradation to the ordinary diameter of the cells of this 
class.—We commonly meet with three or four ranges of 
intermediate cells before coming to those of males; the 
first ranges of which participate in some measure in the 
irregularity of the former. 
But it is upon the construction of the bottoms of the 
intermediate ranges of cells that this variation of their 
architecture chiefly hinges. The bottoms of the regular 
cells of bees are, as you are aware, composed of three 
equal-sized rhomboidal pieces; and the base of a cell on 
one side of the comb is composed of portions of the bases 
of three cells on the other: but the bottoms of the in- 
termediate cells in question (though their orifices are per- 
fectly hexagonal) are composed of four pieces, of which 
two are hexagonal and two rhomboidal; and each, in- 
stead of corresponding with three cells on the opposite 
side, corresponds with four. The size and the shape of 
the four pieces composing the bottom, vary; and these 
mtermediate cells, a little larger than the third part of 
the three opposite cells, comprise in their contour a por- 
tion of the bottom of a fourth cell. Just below the last 
range of cells with regular pyramidal bottoms, are found 
cells with bottoms of four pieces, of which three are very 
large, and one very small, and this last isa rhomb. The 
two rhombs of the transition cells are separated by a con- 
siderable interval; but the two hexagonal pieces are ad- 
jyacent and perfectly alike. A cell lower, we perceive that 
the two rhombs of the bottom are not so unequal: the 
contour of the cell has included a greater portion of the 
