856 
Journal of Agricultural Research 
Vol. XXI, No. II 
several values from their mean is only about io per cent of the mean. 
In dealing with biological material we should not be justified in attach¬ 
ing any significance to such slight differences, since they might just as 
probably be due to factors which produce chance variations. 
It therefore seems evident from these figures that the growth output 
of a mother shoot is practically independent of the number of buds it 
possessed or of the number of laterals it produced. On the mother 
shoots on the pruned trees the largest number of laterals in the various 
classes ranged from 4 to 36, nevertheless the total production of laterals 
was remarkably uniform, considering the variability commonly encoun¬ 
tered in biological material. It should be noted that the differences in 
length of the pruned shoots were due to the varying amounts of wood 
removed when they were pruned and not to inherent differences in their 
size. 
These results seem to correspond with the data obtained by Miss 
Brenchley ( 1 ) on the total dry-weight production of barley and mustard 
in pots containing uniform quantities of soil. She found that, so far as 
total dry weight was concerned, there was no real difference whether 
there were 1, 2, 3, 4, or 5 plants per pot. 
In this inquiry we have so far neglected one question which the reader 
is likely to raise, “Is there any tendency toward compensatory growth 
following the amputation of various amounts of the mother shoots?’* 
In the category of biological ideas there has been some sort of idea that 
an injured organism, or one of its members, has a tendency in its growth 
to “restore lost parts.” This idea, though often expressed in crudely 
anthropomorphic phraseology, has dominated many purely qualitative 
studies in regeneration. It is deserving of study in a quantitative way 
if possible. 
The total new growth plus the length of the mother shoots may be 
compared for the different classes of mother shoots. Table IV gives 
the means of these various sums and the root-mean-square deviation of 
the series from their respective averages. For the pruned shoots the 
values range from 260.0 cm. for the “5” class to 381.7 cm. for the “26” 
class, a difference between extremes of 121.7 cm. The mean total 
lengths of laterals alone (Table III) had an extreme difference of 75.9 
cm. The wider range of values in Table IV is to be attributed to the 
inclusion of the length of the mother shoots. Another way of showing 
the same relation is by the value of the root-mean-square deviation 
which measures the dispersion of the variates from their mean value. 
In Table III the root-mean-square deviations for the pruned and un¬ 
pruned shoots were 9.4 and 11.0 per cent of their respective means. In 
Table IV these deviations were raised to 12.8 and 15.1 per cent of the 
respective means. 
