X, C6 Rundles: Studies in Rice 361 
From Table I it appears that in the different varieties of rice 
there is a wide variation in the number of seeds per cavan. This 
variation is a result of the difference in the size of the individual 
grains and in the awn character. 
The varieties should be classified as to the number of seeds 
per cavan in order to make allowance for the great variation. 
A cavan of palay with seeds of the smaller type may contain 
twice as many seeds as a cavan of the larger type. For this 
reason three groups are established: 
Group A (large). This may include the varieties which have 
large-sized grains or medium-sized ones having the awn char- 
acter or those varieties which require up to 1,650,000 unhulled 
grains to make a cavan. This group is represented in Table II 
with 1,500,000 as an average. 
Group B (medium). This group includes the varieties which 
have medium-sized grains or small grains having the awn char- 
acter. It may include varieties requiring from 1,650,000 to 
2,250,000 grains, to make a cavan. It is represented in Table II 
with 2,100,000 as an average. 
Group C (small). This group includes the varieties which 
are small, requiring more than 2,650,000 grains to make a cavan, 
and is represented in Table II with 2,700,000 as an average. 
Two varieties of palay similar in dimensions but differing in 
the awn character usually fall into different groups—the awned 
requiring the lesser number of grains to make a cavan. (Com- 
pare balalangonon with castafia, Table I.) 
It is important for the farmer to know the approximate num- 
ber of seeds required to make a cavan in order to have some 
knowledge of the number of seeds sown on a given area. A 
little practice in counting enables him to classify varieties. 
In the class room in the College of Agriculture the boys 
counted a definite volume of seeds representing different varieties 
and scattered the seeds at varying distances over a surface of 
a square meter—both broadcast and in rows. The exercise 
trains the eye to note differences in size of grain and variation 
in the rates of sowing. When the seeding rate is excessive, 
that fact is made very evident. 
The number of seedlings given in Table I are estimated upon 
a germination of 80 per cent, thus allowing for a possible loss 
of 20 per cent in the seed bed. 
