ran<;e of broods in order OF FUTl'KE APPEARAND Eg 
33 
The widely separated swarms listed for Illinois and South Carolina 
are extremely doubtful, the latter also being much too far south for the 
17-year race, and both probably are based on confusion of some of the 
annual species of the Cicada with the periodical species. 
The distribution, by States and counties, as now known is as follows: 
Illinois. — Whiteside (f). 
Massachusetts. — Duke. 
New York. — Chautauqua (?). 
Ohio. — Belmont, Carroll, Columbiana, Jefferson, Mahoning, Trnmbnll. 
Pennsylvania. — Armstrong, Beaver, Butler, Cambria, Clarion, Forest, Huntingdon, 
Indiana. Jefferson, Mercer, Snyder. Venango, Washington. 
South Carolina. — Barnwell ( I . 
West Virginia. — Marshall. Ohio. 
Brood XXI. — Sepiendeeim — 1901. (Fig, 7.) 
Iu the main, this brood (XV Walsh-Riley) covers a rather compact 
territory, extending from the southern part of West Virginia across 
Fig. 7.— Map Bhowing distribution of Broods XXT and X. 1901. 
Virginia into North Carolina. A very doubtful record, referred to this 
brood, is given for Ohio, and a record from Loudoun County, in the 
northern part of Virginia, which seems to be authentic, but which may 
have been based on precursors of Brood XX. since it is bo widely sep- 
arated from the main swarm. 
The occurrence of a swarm on Marthas Vineyard in 1833, as recorded 
by Dr. Harris, would indicate the existence of this brood on that island. 
