10 BULLETIN 177, U. S. DEPARTMENT OF AGRICULTURE. 
In 1910 the census returns show the number of farms reporting j 
dairy products as well as the number reporting sales. Thus it is pos- I 
sible to compute the average production as well as the average sales I 
per farm. Such a computation brings out in a striking manner the | 
fact that the average quantity of milk sold per farm of those report- 1 
ing sales was larger than the average quantity produced per farm 1 
of those farms reporting production. This is because the large dairy I 
farms that sell milk constitute a large proportion of those upon whose I 
reports the average sales are computed, while average production is j 
based upon all farms reporting milk production. A similar com- ] 
parison for butter brings out the fact that the average production j 
for farms showing butter production is larger than the average sale I 
for farms reporting sales ; that is, farms of both small and large pro- 
duction sell butter, whereas the farms selling milk consist in large 
proportion of farms of large production. 
DAIRY PRODUCE AS A PRINCD7AL SOURCE OF INCOME. 
In 1900, farms were classified according to the principal sources of 
income. The average number of cows per farm which reported that 
the principal income was from dairy products ranged from 12 per 
farm in the Pacific and Middle Atlantic States to less than 5 in the 
East North Central. 
FARM CONSUMPTION OF BUTTER. 
If we take the total quantity of butter reported as produced on 
farms and subtract from it the quantity reported as sold, we have 
practically the amount of butter consumed on farms. By this method 
it is estimated that both in 1900 and 1910 the annual consumption 
averaged 153 pounds per farm. Whether this is the actual average 
or not can not be proved, but it differs considerably from a similar 
average obtained by another method which will be shown later in the 
bulletin. 
CONSUMPTION OF MILK IN NEW YORK CITY. 
One of the largest factors in determining the consumption of milk 
after it leaves the farm is the. number of people living in cities and 
towns. In order that a clear conception may be obtained of the 
influence of market milk upon butter and cheese production, a study 
of the milk supply of several large cities might be of interest. Prior 
to 1842 practically all of the milk consumed in New York City was 
brought by wagons from the surrounding counties. In 1842, when 
the Erie Railroad was under construction, one of the New York City 
milkmen began to ship milk from Orange County. This milk proved 
to be of such good quality that the traffic spread rapidly. In a few 
years the Harlem division of the New York Central began to haul 
milk from the counties on the west bank of the Hudson River. At 
