New York AGRICULTURAL EXPERIMENT STATION. 331 
the intervening cloth hangs down between the rods within about 
an inch and a half of the bottom of the box. The end of the cloth 
is left sufficiently long to reach the bottom of the box. The seeds 
are placed in the folds between the rods. About half an inch of 
water is poured in the bottom of the box to supply moisture. This 
apparatus is still popular with those who test seeds and is known 
as the Geneva Germinator. It is convenient, durable, and easily 
sterilized to remove any invading germs of decay. 
The question was raised as to the accuracy of such germinators 
in testing the value of seeds, or rather the accuracy of seed test- 
ing; 7. e., whether 50 or 100 seeds selected from a large quantity 
and tested give a correct indication of the value of the lot 
To determine this point duplicate tests were made from the same 
batches of seed and the results compared. In the case of seed 
which was low in vitality either from age or any other cause, the 
results were very variable, ranging from o to 4o per ct. The 
greatest difference found in two tests of seed not more than one 
year old was 21 per ct. and the average difference for seeds, grown 
the previous season, of different varieties of twenty-four different 
kinds of vegetables was only 4.7 per ct. Later this work was 
carried farther to determine if a portion of these discrepancies 
might not be eliminated by increasing the number of seeds. 
With watermelon seed four years old it was found where 50 
seeds were used in each lot tested the “ variation was 20 per ct.; 
with 100 seeds, just 20 per ct.; with 200 seeds, it was 6 per ct.; 
with 300 seeds, it was 5 per cent.; with 400 seeds, it was. 2 per ct.; 
and with 500 seeds it was 6 per ct.’ With onion seeds where 50 
‘seeds were used the “ variation in germination between duplicates 
is expressed by 24 per-ct.; with 100 seeds, by 15 per ct.; with 200 
seeds, by 6 per ct.; with 300 seeds, by 4 per ct.; with 400 seeds, by 
3 .per ct. with soo séeds, by- 2 per cent.” It may thus be seen 
that increasing the number of seed does to a large degree remove 
the error arising from the seeds not running even, that is, not all 
being of a like degree of vitality. 
It was found both in the investigations outlined above and in 
the regular garden work of the horticultural department, that much 
of the seed on the market is defective. This applies to seeds from 
all sources, and while criticism is not made of any individual seéeds- 
man the statement is made that our “experience justifies the re- . 
flection that our seedsmen scarcely exercise the proper care in 
the sending out of their seeds and in assuring quality in their 
