Sept. is, i 9 i6 Dwarf Eggs 985 



the variation constants that it is unfair to compare them with the con- 

 stants for normal Barred Plymouth Rock eggs. 



By means of the data given in Tables II, III, and IV it is possible to 

 compare the size, shape, and degree of variability of the several groups 

 of dwarf eggs both among themselves and with normal eggs. 



A. — SIZE RELATION 1 OF DWARF AND NORMAL EGGS 



The means given in Table IV show mathematically that all classes of 

 dwarf eggs are of lighter weight and both shorter and narrower than 

 normal eggs. This fact is, of course, obvious from the most casual 

 inspection of dwarf and normal eggs. In comparing the different classes 

 of dwarf eggs with each other it is necessary to keep in mind that the 

 number of cylindrical eggs is so small that the means determined may not 

 represent the true means for this class of eggs. Of the eggs studied, 

 however, the mean prolate-spheroidal, or egg-shaped, egg was decidedly 

 heavier than the mean cylindrical egg. It was also decidedly broader. 

 It can be seen from the means given in the table that the mean weight 

 and the mean breadth for both groups of cylindrical eggs are smaller 

 than the mean for the same character for any group of the prolate- 

 spheroidal eggs. The mean lengths for all cylindrical and all prolate- 

 spheroidal eggs may be compared by calculating from the means in 

 Table IV the weighted mean for each of these shape groups. The mean 

 length for the cylindrical eggs is 38.22 and for the prolate-spheroidal 

 eggs it is 36.23 — that is, the cylindrical eggs studied were much lighter 

 in weight, decidedly narrower, but slightly longer than the eggs of the 

 prolate-spheroidal type. 



The number of each class of cylindrical eggs is so small that the com- 

 parisons of the means for the two classes is of very doubtful meaning. A 

 comparison of the means for the several groups of prolate-spheroidal 

 eggs seems to show that those with small yolks average longer, broader, 

 and heavier than those of the other groups, while the means for the 

 dwarf eggs with some yolk not in membrane (free yolk) are slightly 

 higher than for yolkless dwarf eggs. While the number of dwarf eggs 

 in each group of prolate-spheroidal eggs is larger than in the case of cylin- 

 drical dwarf eggs, the actual number is not very large. In order to deter- 

 mine whether or not the above noted differences are greater than those 

 which might arise from errors in sampling, each difference is compared 

 with its probable error. The first section of Table V gives for each 

 physical character measured the deviation in mean with the probable 

 error, and the ratio of the error to the deviation between normal eggs f 



1 Since the constants derived from the 450 eggs laid on the same day are measures of an absolutely random 

 sample of Barred Plymouth Rock eggs, these constants are used in calculating the difference between 

 dwarf and normal eggs in the case of length, breadth, index, and weight. Data on the weight of the egg 

 parts were not taken on the 450 egg series. Therefore the only available constants for these characters are 

 those determined from all of the first-year eggs of the small flock. 



