28 REpoRT OF DEPARTMENT OF BACTERIOLOGY OF THE 
INTRODUCTION. 
Alfalfa has long been grown in the central part of this State, 
notably around Syracuse, but within the past ten years its culture 
has spread. Now, alfalfa is being successfully grown in every 
agricultural county of the State. The main reason why the growth 
of this valuable forage plant was so long restricted to a few locali- 
ties is found in the particular requirements of this plant; and the 
recent wide spread of its culture is evidence that these requirements 
are being fulfilled. However, failures are still numerous and the 
factors which make for success are not always understood. Good, 
clean seed is a prime requisite, the soil and subsoil should be well 
drained, the soil should contain a fair amount of fertility and should 
be so fitted as to prepare a good seed bed and at the same time 
kill a maximum number of weeds. 
When proper attention has been given to all of these things there 
still remain the important factors of inoculation with the bacteria 
which enable alfalfa to obtain nitrogen from the air, and of lime. 
This bulletin presents the results of a study of the influence of in- 
oculation as shown in cooperative experiments conducted during the 
past three seasons on sixty-seven farms in this State. In these ex- 
periments the aim was to ascertain the facts as to the need of 
inoculation in growing alfalfa on these farms. As this need is im- 
perative in many cases, attention has been given to different means 
of supplying inoculation. 
While these experiments have the advantage that they embody 
practically all types of the agricultural conditions to be met with in 
the State they are open to the criticism that they were rarely con- 
ducted by experienced experimenters and our supervision was limited 
to directions through correspondence and an occasional visit. These 
visits gave an opportunity to verify practically all of the data fur- 
nished by the farmers except that concerning the comparative yields 
of the respective plats and in this particular the farmers were well 
qualified to determine and report the facts. 
In work of this kind the need of a check plat to serve as a basis 
for computing the effect of any line of treatment is imperative. 
The most common error with these experimenters was either to 
omit the check plat or to so locate it that it was rendered worthless 
by surface drainage or agricultural operations. All experiments 
where the check plat was absent or where, for any reason, the re- 
sults from it were of doubtful value, have been eliminated from our 
computations. 
The peculiar feature of this work was the attempt to determine 
