1913] ELLIS—SEED PRODUCTION IN YUCCA Ps 
the average number of perfect seeds or more. That the presence of 
a number of larvae in a single pod need not prohibit average seed 
production is shown by plants 2 and 9. A pod of the former pro- 
duced 172 perfect seeds, although 12 larvae had been in the pod. 
The pod of plant 9 also produced good seed in excess of the average 
even though there were 5 larvae in it. The number of larvae pres- 
ent did not seem to have any effect on the total number of seeds 
(both infertile and perfect) produced. 
The production of the average number of perfect seeds by pods 
with many larvae calls attention to the influence of parasites of 
Pronuba. Of the 30 pods examined, 7 were found without injured 
seeds. In several other cases, although some of the seeds had been 
eaten, the dried skin of the larva was found in the burrow in the 
seed column where it had died. It is well known that most species 
of insects are parasitized by other insects, particularly by minute Hy- 
menoptera. These parasites are usually parasites of the egg or the 
larva of the host. The action of such parasites would account for 
the death of the eggs and young larvae of Pronuba. The result of 
the presence of these parasites would be advantageous to Yucca, 
for once the parent Pronuba has pollinated the blossom of Yucca 
its value to that Yucca plant ceases. The destruction of seeds 
vy larvae is unnecessary, as proven by 7 cases, and, so far as the 
particular plant is concerned, disadvantageous. The advantage 
comes only to the species as a whole, in supplying food for the . 
Pronuba larva. The elimination of some of the eggs and larvae of 
Pronuba by parasites could also be of value to Yucca as a group, 
if not carried too far. These egg and larva parasites aid in pre- 
serving the balance by reducing the number of larvae toward the 
optimum for the plant, thus saving the plant the seeds eaten by the 
extra larvae. If the 7 pods which lost no seeds as the result of injury 
by larvae be dropped, the average number of seeds eaten per larva 
is raised from 19 to 24, and the number lost per pod from 58 to 72. 
The destruction of the egg in these 7 cases makes this important 
change in the averages, and if the saving of seed to the plant by the 
destruction of the immature larvae could in some way be figured in, 
the change would be still greater. Of course these eggs and larvae 
may not have been destroyed by parasites, although that is rather 
