THE BROOD-REARING CYCLE OF THE HONEYBEE 5 
the location of the frame in the hive body, and the particular side 
of the frame. A net of wires forming 1-inch squares is permanently 
fixed to the holder at such a point as barely to clear the surface of 
any brood frame in the holder and still be in focus. The squares, 
showing clearly in the negative (Plate I, B), divide the brood area 
into such small sections as to render possible an extremely accurate, 
direct count. 
For recording the counts from each negative, a card is ruled into 
squares identical in size and number with those in the picture itself. 
All squares corresponding to areas containing only sealed worker 
cells may be credited with the number of cells contained in 1 square 
inch; but in squares containing unsealed as well as sealed cells, the 
number of unsealed cells must be deducted first. There are many 
contradictory statements as to the number of cells per square inch, due 
in part to attemps to derive it mathematically from the dimensions 
of some single cell instead of counting the actual numbers in areas 
large enough to get a trustworthy average. Watson, by making 
such counts, found the average number per square inch to be slightly 
in excess of 26. This number has been used in the results presented 
here. Some variation exists between individual combs, however, 
possibly due to the foundation used. Thus, in combs from certain 
foundation the writer has found 26.3 worker cells per inch, and in 
combs from foundation bought after the results for 1921 were ob- 
tained he has found approximately 27 worker cells per square inch. 
Much of this latter type of foundation has been used since that date, 
and subsequent results are therefore being calculated on this basis. 
It is very evident that the general relations of a curve based on the 
amount of brood counted will remain the same, regardless of whether 
in the counts the number of cells per square inch is taken to be 26, 
27, or some other figure of nearly the same size. As a matter of fact, 
the difference between 26 and 27 is less than 4 per cent, or less than 
40 in every 1,000. 
In areas containing 50 per cent or more of unsealed cells it has been 
found preferable to count each individual sealed cell. Sealed drone 
cells also are counted individually. They have little influence on 
the totals, however, because by proper selection of brood combs it 
has been possible to keep the total of sealed drone cells well below 
600 on any one count. Since in counting individual cells, either 
sealed or unsealed, some arbitrary rule must be followed in crediting 
them to a particular square; all such cells are credited to, or deducted 
from, the total of the square immediately to the left of the vertical 
dividing line, or below the horizontal dividing line, as the case may 
be. Totals for each card and colony are calculated on an adding 
machine. The photographic record once obtained, the actual count- 
ing may be delayed until any convenient time, so long as the nega- 
tive does not deteriorate. A series of such records permits making 
a year-by-year comparison of any portion of the sealed-brood area. 
Cut films, 5 by 7 inches in size, are used exclusively in this work. 
Before photographing, all of the frames containing sealed brood 
in any given hive body are shaken or brushed free of adhering bees, 
placed in an empty hive body, and immediately carried to the build- 
ing where the photographs are made. The exposures can be made 
in 10 seconds, thus keeping the frames out of the hive for an exceed- 
ingly short time. Danger of chilling the brood is thus reduced to a 
