930 MR. I. H. BURKILL ON SOME VARIATIONS IN 
correctly stated is proved by Table VII., in which the facts are 
summed up. In the last three columns, lettered z, y, and z, the 
percentages of the various stamens which should be present 
according to these two rules are given; while in the other 
columns the numbers and percentages actually found are stated. 
Columns X, Y, and Z correspond with the columns z, y, and z; 
and it will be noticed that the discrepancy between the real 
percentages and those found by these two rules is in no case as 
great as 5 per cent. The constancy of the three stamens of the 
inner whorl superposed to the inner sepals is most marked, 1:03 
being the greatest difference between the figures in columns Z 
and z. With regard to other stamens, it will be noticed that 
frequently the percentage of the stamens of the outer whorl is 
upwards of 5 per cent. greater than the theoretical figure of 
column z, while that of the stamens superposed to the outer 
sepals is the same amount toosmall. Then to this slight extent 
the stamens of the outer whorl have usurped the place which 
theoretically should belong to the two outer-sepaline stamens ; 
in other words, the outer whorl does not exhibit so strong a 
tendency to abortion that it entirely disappears before abortion 
begins in the inner whorl. 
It may be mentioned here that of the five cases in which a 
doubling of a stamen was observed, in four of these cases it 
occurred in the inner-sepaline stamens, once in the petaline 
stamens, and never in the outer-sepaline stamens. 
With regard to the petaline stamens, that marked < in the 
accompanying woodcut is the one most frequently developed. 
Of the other four (£), none can be said to be much more frequent 
than the others. 
Diagram showing position of stamens with regard to the sepals and ovary. 
Petals not inserted. The stamens of the petaline whorl are indicated by 
the letters æ and f, those of the sepaline whorl usually absent are shaded. 
Changes in the Carpels.—The number of carpels entering into 
the ovary may be regarded as constant. Out of the 5697 flowers 
