264 THE INDIAN JOURNAL OF VETERINARY SCIENCE AND ANIMAL HUSBANDRY [I, IV.

formed where a dry product was aimed at. The loss of dry matter in this pit was
found to be 4.85 per cent. and 6.94 per cent. respectively on the corrected and
uncorrected basis (Table VII). This is the lowest loss recorded in any of the pits.
It seems probable that the actual loss was very much greater, but has been
replaced to a considerable extent by material washed down from the spoilt surface
layer where extensive decomposition took place. Further evidence in support of
this view will be brought forward below (see nitrogen losses). In spite of this
large replacement the organic matter loss comes to 10.0 per cent. (Table XV),
which is indeed a high figure, even as it is. The figures of this pit are of interest
in showing what may occur in a flooded pit.

Considering the evidence of the unspoilt pits, it may be said that rapid
filling causes a larger loss of dry matter than slow filling, while watering gives a
better aroma and consistency to the resulting product. So, to obtain the best
results, it is indicated that filling should be done reasonably slowly, and water
should be added in sufficient amount to avoid the production of a very dry silage
and to improve the aroma.

The average loss of the dry matter in this experiment works out to 7.67 per
cent. if all the bags in all the pits are considered. This experiment shows a higher
loss than the last one ( l. c.) which gave a loss of only 4.2 per cent. for dry matter.

The loss of 7.67 per cent. noted here agrees remarkably well with the loss of
7.7 per cent. noted by Amos and Woodman ( Journal of Agricultural Science, XV,
344) at Cambridge in the production of "Acid Brown" silage, though working
with different material, as well as with the loss of 7.5 per cent. noted by Annett
and Aiyer (Memoirs of the Department of Agriculture in India, Chemical Series,
Volume VIII, No. 10), also working with Jowar.

In this connection it is very important to remember that there are two types
of losses in the process of the preparation and utilisation of silage. The first is the
loss resulting from the activity of enzymes, bacteria, etc., which cause the fermen-
tation, and produce the typical characteristics of silage. It is this loss that has
been studied in these experiments and it is only this loss that has been recorded.
The second kind of loss is due to the production of mouldy, rotten silage at the
sides and top of the silo. This, of course, cannot be avoided entirely, but could
be reduced to a minimum by suitable means. It is under this latter category that
the rapidly filled pits, and more especially pit No. 4, show considerably larger
losses than the others (Table III). In pit No. 4, as has already been noted, the
thick layer of rotten silage at the top formed a very large proportion, about a third,
of the total. In this respect also, therefore, the slowly filled pits score an
advantage over the rapidly filled ones.