232 
R. H. Dastur and W. T. Saxton 
as the main shoot and the normal branches die, and the lateral roots 
then function as the lateral roots of the accessory system. 
Accessory bundles are also found at a higher level in the shoots, 
both normal and reproductive, but they arise quite independently 
of the accessory system we have described, and are small regular 
and inconspicuous externally. These are incorrectly described in a 
recent paper by Mr Sabnis(i) as anomalous cortical bundles, but as 
they arise inside the endodermis it is obvious that the word “ cortical ” 
is a misnomer. 
Discussion 
The observations described in this paper make it clear that the 
whole development of the accessory bundle system is directed to the 
gradual separation of reproductive branches which will eventually 
become established as separate plants. So far as we are aware, such 
a provision for vegetative reproduction is markedly different from 
anything previously described. The plant is a prominent member of 
a psammophytic association in which a likelihood of partial burying 
in the sand clearly exists, and though we have not actually seen the 
process taking place, our observations point to this conclusion, since 
much diversity exists in the depth at which the transition region is met 
with in plants of the same age. Such a state of affairs will obviously 
accelerate the separation of branches initiated by the process we have 
described. We thus see that when the main axis of the plant perishes, 
a circle of branches separated to below the ground level is already 
established, and it cannot be doubted that the separation must 
eventually become complete. 
The family Leguminosae is well known to be remarkable for the 
frequency with which abnormal or unusual vascular structure is met 
with, but none of the accounts of such structures indicate, so far 
as we are acquainted with them, that they are connected with any 
specialised method of vegetative reproduction, nor do we think, 
regarded purely as anatomical peculiarities, that the structures 
hitherto recorded are quite comparable with those here described. 
Further reference to literature would therefore appear superfluous. 
It is worthy of notice that though the plant is seen flowering 
during the major part of the year it seldom seems to develop seeds. 
This season, however (November, 1920), the mature pods have been 
collected in large numbers, a phenomenon which we attribute to the 
almost complete failure of the latter half of the monsoon. 
