1912] HARRIS—STAPHYLEA 399 
room which it demands is that necessary for its attachment; the 
pressure of adjoining seeds is a factor entirely removed. Staphylea 
has the advantage that the number of seeds developing is so small 
that it is easy to select material in which the stretching of the pod 
by a crowding of the seeds can hardly be a possibility. For pods 
with one to three seeds, there may be only a single seed in each 
TABLE XIII 
1906 1907 
CLass 
r Mean length f Mean length 
t Seed 
FOO lS anaes g60 5-57 519 6.26 
2 Seeds 
990 6 ea ee es 291 6.390 227 6.83 
2-0-0 io See aos 204 6.46 105 6.96 
3 Seeds 
The 8 a eee 7.23 32 -19 
SiO) oe ee 172 6.07 117 7.21 
POO. Ce ees 47 7.2% 6 6.96 
4 Seeds 
rae, foe Oe GE? ora en 64 7.56 40 7.38 
PPO cls ee a 23 7-39 a) 7-33 
=, eg le? Cet eer eC ee 30 7.90 23 7.61 
Mote ere ee 14 7.64 5 7.80 
locule, that is, seed formulae: 1-0-0, I-I-0, I-I-I. When more 
than three are produced, one of the locules must have two or more. 
Now the problem is simply this; given fruits producing the same 
number of seeds, are those in which two or more of the seeds are 
produced in the same locule larger than those in which but a single 
seed is produced in each locule ? 
Table XIII gives the number of pods available and the mean 
length of pod for the chief seed-formulae (including about 91 per cent 
of the whole number of fruits available) for the 1906 and 1907 series. 
The results are also shown graphically in fig. 5. Here the solid 
dots give the empirical means for the individual seed-formulae for 
1906 and the circles those for 1907. The vertical lines (solid for 
1906 and broken for 1907) indicate the mean length of pods with 
I, 2, 3, and 4 seeds per pod, irrespective of the distribution of these 
seeds among the locules. Certainly there is no clear evidence in 
this figure that pods in which the seeds are grouped in less than 
