110 
Improvement of the Plants of the Farm. 
the change of seed from one district to another is generally 
regarded as advantageous, but the soundness of the opinion has 
not been tested by exact experiments. 
Size of the Seeds of Cereals. — Mr. Darwin informs us that 
Sir J. Hanmer, writing of plants for the flower-garden in 1660, 
says that, in " choosing seed, the best seed is the most weighty, 
and is had from the lustiest and most vigorous stems." M. 
Henry Vilmorin gives the best possible and the briefest answer 
to my question, " Is the heaviest sample considered the best for 
sowing ? " when he replies, " The heaviest, not the plumpest." 
The distinction is of great importance. An experimenter assures 
me that he pruned the seed-stems of a turnip, and thus ob- 
tained seeds of unusual size, which proved very inferior in 
vitality to the seeds of common size. All seeds whose size is 
thus increased by overfeeding, such as the seeds of wheat planted 
thinly in the ground, or on land unusually rich, will probably 
be lighter in proportion to their size, and deficient in constitu- 
tional power. It is quite possible that large seeds, though really 
inferior, may be sown sometimes with success in genial years 
in kind land, or in cases where over-seeding is customary, 
because large-sized grain gives less seed per bushel, but that 
does not affect the general rule that good wheat should be of 
high specific gravity according to its kind. INIany of my in- 
formants do not object to small sound seeds of any kind, but a 
sample of small seed is quite different from the small seed of 
a sample. I have received sanguine narratives of the increased 
produce of barley when the small-sized grain had been carefully 
sifted out. But this cannot, in the face of other evidence, be 
regarded as a universal method of improving plants. 
Over-size seeds are always inferior, although no doubt the 
seed inherits characters which are in some degree dependent on 
size, and large seed contains — other things being equal — more of 
the desirable qualities than small seed, and these are more 
likely to be propagated and handed down b}- such seeds, provided 
they have no defect. It is recognised, however, that very large 
seed produces a coarse growth. A seed -grower sifted out the 
very large seeds of mangolds and saved the produce for seed. 
Not only were the plants thus produced coarser than usual, but 
the seedsman for whom he had long grown this mangold, though 
unaware of the experiment which had been tried, detected the 
degeneracy of the stock. 
The vigour of the germ of seeds is not always in proportion 
to the beauty or plumpness of the grain. Wlieat from an inferior 
soil is usually preferred for sowing, and M. Vilmorin remarks 
that the best agriculturists of French Flanders, having excellent 
