1924] : 
HARRIS—THE INFLORESCENCE OF MANFREDA VIRGINICA 425 
The great difference in the mean number of flowers and fruits 
renders a discussion of the comparison of variability impossible 
except on the basis of constants involving correction for mean 
numbers. These constants are afforded by the coefficients of 
variation which show that the variation for number of fruits per 
inflorescence is regularly higher than that for number of flowers 
per inflorescence. Тһе coefficients of variation for number of 
flowers range from 28.3 to 33.3, whereas those for number of 
fruits range from 43.3 to 506.3. 
This result for relative variability in flower and fruit number is 
substantiated by results for Crinum longifolium and Celastrus 
scandens, but not for Staphylea trifolia. Тһе actual coefficients 
are as follows: 
Flowers Fruits 
Crinum longifolium (Harris, 1912 а)................ 22.80-24.32 | 31.73-33.70 
elastrus scandens (Harris, 1909)................... 26.80 48.58 
Staphylea trifolia (Harris, 1909 а).................. 64.44 53.04 
In a series of inflorescences of varying numbers of flowers per 
inflorescence the distribution of the position of the flowers on the 
inflorescence must be considered in relation to the number of 
flowers in each position, the ovaries of which might have developed 
into capsules. 
Table rx shows the number of flowers formed and the number 
of capsules matured in the various positions on the inflorescence 
for З of the series of data.! From these data the percentage fre- 
quencies of numbers of capsules matured in each position on the 
inflorescence have been derived and are represented in fig. 5. 
The higher positions, of course, occur only on the larger inflores- 
cences. In any sample which it is practicable to secure, the 
numbers of flowers and fruits in the more distal positions are во 
small that percentages caleulated upon them are very irregular. 
Only percentages based upon at least 100 flowers have been in- 
cluded in the diagram. 
1 The reader who cares to do so may determine the number of flowers and number 
of fruits for the 3 series of inflorescences in which number of flowers and number of 
fruits only was determined by subtracting the appropriate sections of table x from 
table 1x. 
