NO. 2192. DRAG0NFLIE8, CALIFORNIA AND NEVADA— KENNEDY. 517 



swiftness to become warm and muddy. In the steep and narrow 

 mountain gorges where the rushing torrents pour down through the 

 shade of the redwoods and alders, this dragonfly adds a note of mys- 

 tery to the scene, for the individuals with their strange ophidian 

 coloration gUde noiselessly up stream or down, never showing that 

 curiosity toward strangers or unusual surroundings which is exhibited 

 by the libeUuhnes of the sunny valleys, but always moving straight 

 ahead as though di-awn irresistibly onward. Only males are common 

 on the streams, the females seldom resorting to the water except to 

 oviposit. The males, as indicated above, fly on the longest beats I 

 have observed for any dragonfly, for they fly continuously up stream 

 or down until they come to the head of the stream or to the slow water 

 below, or until some unusual obstruction turns them aside, when they 

 face about and fly as steadily in the opposite direction. The course 

 is usually a foot or two above the surface of the stream and goes 

 through dense shade and any loose brush or foliage which may hang 

 over the water. OctogompJius, with the possible exception of Aeshna 

 walJceri, is the only other western dragonfly which wiU fly in such 

 close dark places. Because of this habit of flying in long beats this 

 dragonfly is not easily taken, as the collector has but a single chance 

 at each individual. 



I have seen but one female over the water. This one w^as oviposit- 

 ing. I have seen two others wliich I thought were females on the 

 mountain side several hundred feet above the stream. As I have 

 seen many males on the streams and only the single female, I conclude 

 that these seldom resort to the water except to oviposit. 



In the streams of the Coast Mountains of California, v>^here I have 

 had opportunity to observe the habits of Cordulegaster most, it shows 

 a marked upstream migration of the imagoes. The eggs are laid in the 

 shallow water along the edges of the stream and the nymphs wander 

 aimlessly over the bed. Because the nymphs are free on the stream 

 bottom each freshet dm-ing the three or four years of nymphal life 

 serves to wash them farther dowTi stream so that when they come to 

 emerge they may be far downstream from the place where the eggs 

 were deposited. On Stevens Creek (Santa Clara County, California) 

 exuviae were abundant almost down to the Trout Farm, while few ima- 

 goes were seen below Soda Spring, which is 2 miles farther up the 

 creek, and imagoes were common on the divide at the head of the 

 creek, where few exuviae were found. Exuviae were abundant on the 

 lower part of Zyante Creek (Santa Cruz County, CaHfornia), where 

 no adults were seen, though it was at the height of the Cordulegaster 

 season. From the above and similar observations on Octogomplius, 

 I have concluded that the imagoes show such a preference for the 

 swifter water of the upper reaches of these streams that when they 

 reach the upper part of the stream in their first flights they remain 



