cies of swallows, most notably bank and barn swal- 
lows, did the same thing. 
Observers generally give departure dates for the 
cliff swallow in mid-September or late September 
(Oberholser 1917, and Schafer 1917-1918) , and our 
censuses also showed that most of the cliff swallow 
population is gone from Illinois by the end of Septem- 
ber. A few birds have been recorded as late as October 
99 in the south and October 24 in the north (Fig. 24; 
Clark & Nice 1950). Frank Bellrose (personal com- 
munication) observed an exceptionally late aggrega- 
tion (“thousands”) on October 16, 1970, flying south 
along the Illinois River Valley near Havana. 
As with other swallows, cliff swallows were seen in 
greatest numbers near the large rivers, and on the 
western side of the state (Fig. 24). Cliff swallows 
were numerous along the Rock River, as well as the 
Mississippi, whereas tree swallows were found in 
much smaller numbers along the Rock River than 
the Mississippi. 
Because of the exceptionally large spring migration 
in 1968, when we censused in northwestern Illinois, 
the spring-fall ratio of cliff swallows for the region 
(2 birds in the spring to | in the fall) differed greatly 
from the ratios in central (1 bird in the spring to 16 
in the fall) and southern Illinois (1 in the spring to 
7 in the fall). In most years, the fall population prob- 
ably far exceeds the spring population throughout the 
state. 
Food Habits 
Forbes (1878, 1882) examined the stomachs of six 
cliff swallow specimens, one of which was collected in 
Tazewell County in the summer. He gave no locality 
or season for the other specimens. Beetles, including a 
scavenger beetle (Aphodius inquinatus), were the dom- 
inant food items, with Hymenoptera (wasps and ants) 
next in abundance. 
As in the case of other species of swallows, we have 
noted that the large aggregations of migrant cliff swal- 
lows coincide with conspicuous populations of midges 
(Chironomidae) . In May, 1968, we observed cliff swal- 
lows feeding steadily on chironomids, notably Chir- 
onomus crassicaudatus. 
PURPLE MARTIN (Progne subis) 
(Fig. 26, 27, and 28) 
Spring Migration 
Cooke (1888) hinted that the purple martin’s mi- 
gration 1s nocturnal, but there is also some evidence of 
diurnal migration, as discussed below under Fall 
Migration. Martins have been observed to arrive at 
their nest sites at various times of the day, morning 
and afternoon (Gault, unpublished notes 1885-1905, 
Cooke 1888, and Cooke & Widmann 1884), and Wid- 
28 
mann observed a few birds in what was apparently 
northward migration at 5:30 P.M. on March 30, 1885 
(Cooke 1888). In view of the popularity of this spe- 
cies and its tendency toward gregariousness, it is sur- 
prising that no one has reported a large scale migra- 
tion of martins in Illinois. 
The earliest arrival date reported for the martin in 
Illinois was February 24, 1923, at Metropolis (Conley 
1926) , an exceptionally early arrival. Alleged martin 
records for February in Illinois often turn out to be 
based on misidentifications, but probably there ar 
reliable (though rare) records in this month. Marcl 
records prior to the 10th are probably unusual even ir 
southern Illinois (Johnston & Hardy 1962), anc 
though there are reports as early as March 5 for cen 
tral Illinois (Nolan 1956) , and March 10 for the nortl 
(Boulton & Pitelka 1937), the very first martins, usu 
ally adult males, are not likely to be detected befor 
March 15 in the south, and March 23 in central ani 
northern Illinois. Martins are subject to stress an 
mortality from cold (Jones 1910, Kimball 1889, Smit 
1943, and Musselman 1966) though they have bee 
known to survive 2 or 3 days of severe cold (15° F. 
and snow (Nolan 1956, and Schafer 1921). Most ¢ 
the state population does not appear until April ¢ 
May. Yearlings arrive later, often weeks later, tha 
adults (Cooke & Widmann 1884). The migration 
prolonged, regularly extending from March ini 
June, and our peak counts came between mid-Apr 
and the end of May (Fig. 29). The counts were lor 
est in southern Illinois and highest in the north. 
Distribution 
The general distribution of the purple martin: 
shown in Fig. 28. Though martins almost certain 
nest in every township in the state, the Illinois nesti! 
distribution (Fig. 30) has not been recorded in s! 
ficient detail; there are no published records for whc 
counties of the state. 
Illinois citizens are showing increasing interest | 
attracting martins, in part because of the public 
given the species by the Griggsville Wild Bird Socie 
Griggsville, Illinois is called the Purple Martin Ci 
ital, and is the site of an industry based on the bi 
(the manufacture of nest boxes and auxiliary equ’ 
ment) , and of a news publication, “The Purple M 
tin Capital News,” an interesting source of infort’ 
tion on martins and other birds. 
Nesting Habitats and Populations 
Even from early historical times purple mart 
have rarely been found nesting in natural cavities’ 
Illinois (Ridgway 1874, and Ries & Werner 1946) . 
notable recent exception is the record of Peter 
(1966), who found martins nesting in tree stur 
standing in the water of the Mississippi River yy 
Thomson (Fawks 1966). There are no records of “! 
