DETERMINATION OF DIGESTIBILITY CO-EFFICIENTS             305

The most noteworthy points about these figures are the values of the protein
on one side and the ether extract on the other. When the amount of concentrate
was least, viz., 1 lb., the protein digestibility of paddy straw worked out at a higher
and positive figure, viz., 5.91 per cent., whereas with increased amounts of cake the
figures were not only lower but definitely negative. This is suggestive of the fact
that a higher value was assigned to the share of cake, leaving a low or negative value
in favour of paddy straw. In the case of the ether extract the position was re-
versed. With increasing doses of cake the digestibility has definitely increased,
viz., 48.6 per cent., 53.24 per cent. and 59.37 per cent. in the 1 lb., 2 lb. and 3 lb.
groups respectively. Or in other words a lower value in the case of ether extract
was assigned for linseed cake, thereby leaving a higher margin in favour of paddy
straw. That this was probably the most likely cause is borne out by the figures
of cake digestibilities obtained by the graphical method as well as by multiple
regression equations. In the following table the results obtained by all the
methods are given.

                                 TABLE V.

                   Digestibility of linseed cake.

It will be noted that whether according to the worked-out figures of paddy
straw (Table IV) or on a comparison with the figures by the new methods (Table V)
the assumed cake digestibility figures of Henry and Morrison appear definitely
higher for all the components except that of ether extract in which case it is consi-
derably lower. These have been correspondingly reflected in a progressive rise
and fall in almost all the figures of paddy straw under the 3 sets of 1 lb., 2 lb. and
3 lb. combinations (Table IV), the greatest difference occurring in the cases of ether
extract and crude protein, the latter giving also negative values. Such differences
with protein in particular seem to have also been obtained by Warth and Gossip
[ 1928 ] and Warth [ 1923 ] as are shown in the following table.

* This is the only value which has not been found satisfactory. The cause of it has been discussed
later.