12 [ ASSEMBLY 
system, so called, unless this plat system be interpreted in accord- 
ance with scientifically settled principles, which must first be ascer- 
tained and valued. 
The trial of germinations, as published, must be considered as a 
portion of the study into the value of figures obtained in their rela- 
tion to duplicates. We have in this trial exact conditions of expos- 
ure of seeds counted from the same package. It is evident that we 
cannot at present interpret the result of an ordinary germination per- 
centage any more closely than these results indicate, as variability 
occurs in nearly every case, and the limits are not as yet exactly de- 
fined. It will be noticed that, as in lettuce seed the variation between 
duplicates averages about four per cent,and the possible variation may 
be even much more, between any two samples which differ in the 
germination but four per cent, we cannot justly say that one sample 
is better than theother. This figure varies also with different species 
of vegetables, and hence each must be studied byitself. A variation, 
which may be called mathematical, also occurs, and. must be con- 
sidered. If we have but ten seeds under trial, the germination result 
can only be expressed as 0, 10, 20, 80, 40, 50, 60, 70, 80, 90 and 100 
per cent, or not within the ten units; if twenty-five’seeds be used, 
we can express our results only by differences of 4; if fifty seeds be 
used, only by differences of two; if 100 seeds, by differences of one. 
Thus to compare two samples — ten seeds of one and 100 seeds of 
the other — we have not real duplicates, for this factor of number 
constitutes a difference. It is probable that for each series of dupli- 
cate germinations there is a certain number which should be used as 
a unit, asin general the differences between duplicates diminish in 
proportion to the number of seed used, but the determination of this 
number must be the subject of experiment, and must be discussed in 
part under the law of probability before we can obtain a constant for 
use. In such a simple matter, apparently, as germination trials, we 
have evidence of the need of careful and exact knowledge. We are 
required, for perfectness of results, to know the value of changing as - 
compared to even temperature, of the influence of the stage of ripe- 
ness of the seed, of methods of gathering and keeping, of the place 
of growth, etc., etc. These points are, however, only referred to here 
for the sake of illustration. 
The subject of nomenclature is a very important one, and it is sur- 
prising that so much has been already accomplished by the public 
without the advantages which an exact nomenclature offers. There 
is scarcely any work we can do toward furthering progress that is of 
greater promise than this. So long as two distant investigators can- 
not be sure that they are experimenting upon the same kind of plant, 
their results cannot be strictly comparable. Even professional 
botanists have not recognized the importance of variety as a factor 
in their experiments. It is unquestionable but that there is a variety 
difference between vegetables of the same species, that results caleu- 
lated for the Dwarf Golden Pop Corn, for instance, which grows 
