- 29 
yield per acre only about half as much as the sands of Holland and. 
- cause weeds the sooner smother the wheat in the absence of cultiva- 
soils, and hints at rewards of skill and ele science ‘a ee pine ‘ 
ishment of negligent and superficial cultivation, One year with — 
another, statistics show that the new lands of America and Australia 4 
" 
the English lands that have had centuries of cultivation, and. that ‘2 
the richer the soil the swifter the decline im yield, in some cases, bes : 
tion. Inaccord with this fact, census and crop reports show a better é i 
average yield in the principal winter-wheat regions than in thespring- Be 
wheat districts, in the old lands than in the new, in the domain of -. he 
rotation and fertilization than in that of primitive wheat-growing. - 
The effective force of this division usually includes about sixty — 
‘persons in the Department, over twenty-three hundred county c¢orre- 
spondents, each having at least three assistants, and State agents, — 
with some thousands of their reporters. Our foreign service is 
under the direction, as heretofore, of Mr. E. J. Moffat, the deputy ‘ 
consul-general at London, who has authority to employ the consular 
resources of this Government in Europe, and the aid of Kuropean a 
‘ 
statisticians and experts in agricultural and commercial informa- 
tion, for the purposes of his.statistical investigation, which relates 
‘primarily to those products in which the United States may have ‘a Ee 
competitive interest. 
The printed reports of the division for the year include an anid if 
report and eleven special reports, making about eight hundred pages. 
The miscellaneous manuscript reports, statements, and letters com- 
prise thousands of manuscript pages. 
The crop-reporting system of the Deparnmices is not a census, 
which would be an impossibility. The units of a crop can not be 
counted till grown, and a veritable count would be useless because 
too late. Crop reporting is counting.in advance by instantaneous ~ 
generalization. . Is it worthless, assome pretend? Fortunately there 
are a few tests available. There is one crop, cotton, that is enumer- 
ated quite accurately. The record of the National Cotton Exchange 
furnished in September, 1887, of the actual growth of 1886 was_ 
6,443,000 bales. The estimate of the Statistician of this Department 
was just 17,000 bales more, and that of the exchange 40,000 bales» 
less. Either is a wonderfully accurate forecast. The Department . 
estimate exceeded the recorded yield by seven ounces of lint per-acre. 
For three years past the forecast has been nearly as close.’ The 
wheat estimates for six years make an average of 439,274,270 bushels; 
the consumption, on the basis of 4% bushels per capita, the actual 
exportation, and the seed, together required 440,735,346 bushels, 
showing an excess over estimates of nearly a fourth of a million — 
bushels per annum. It may not be possible always to make so close: 
a forecast; in the minor crops discrepancies occur; but the continued 
and pronounced success in the principal crops shows what can be ae- 
