SOURCES OF ERROR IN THE OLD RECORDS. il 
In the case of the broods of the 17-year race, the following extend 
on their southern boundaries into the territory of the 13-year race, 
and hence the records of the southern localities are open to some 
question: Broods VI, X, XIV, XVI, I, IV, to a slight extent also 
in the case of Broods IT and III, and doubtfully in the case of Brood 
IX, the possibility of confusion in this last brood depending on the 
accuracy of the extreme northeastern extension of the 13-year Brood 
XIX. 
The following broods of the 13-year race extend northward into the 
territory occupied by the 17-year race, and hence are open to some 
question: Broods XXIII, XVIII, XIX, and XX. 
The records can not be questioned on this ground of the 17-year 
Broods VII, VIII, XI, XIII, and V, and of the 13-year Broods XXIV, 
XXI, and XXII, because these broods are limited in distribution 
to the territory of a single race. 
The most notable instance of the overlapping and consequent prob- 
able confusion of the records is seen in the case of Brood X of the 
17-year race with Broods XXIII and XIX of the 13-year race. The 
remarkable feature in the distribution of the broods named is the 
notable extension northward in Illinois and Missouri of the 13-year 
Broods XXIII and XIX, which fills almost exactly a district which 
would naturally be supposed to belong to the 17-year race and prob- 
ably to Brood X. As pointed out in Bulletin 14, page 26, this cireum- 
stance had special significance in view of the fact that the northward 
extension of the 13-year race is based on Broods XIX and XXIII, 
and that the records prior to 1898 of the former were collected for the 
‘most part in 1868, when this brood was in conjunction with Brood X, 
and of the latter in 1885, when Brood XXIII was also in conjunction 
with Brood X, the limits of which, curiously enough, stop rather 
suddenly at or near the eastern State line of Illinois. The possi- 
bility was immediately suggested that the northern localities assigned 
to Broods XIX and XXIII properly belong to Brood X. 
The occasion of the reappearance of the 13-year Brood XXIII in 
1898 without any important 17-year brood to confuse the records 
and of the 17-year Brood X in 1902, also without a joint occurrence 
of any important 13-year brood, gave the opportunity wished for to 
determine the validity of old records and to fix more accurately the 
distribution of the three broods concerned. 
A very thorough canvass was made in 1898 of the territory covered 
by Brood XXIII, and especially the territory in doubt, by calling 
into requisition the very numerous county correspondents of the 
Statistical Division of the Department of Agriculture and also of the 
Weather Service in addition to the regular correspondents of the 
Division of Entomology. Several thousand replies were received, 
