66 Unwersity of California Publications in Zoology (Vou. 30

On June 23 three nests were found in one marsh. Two of these were
about twenty feet apart, the other some two hundred yards away.
The two closely adjoining nests when found held, respectively, two
and four eggs, the other, eight. The nests with two and four eggs, on
July 2 held, respectively, nine and ten.

The three nests were similarly placed, each in a tussock of long
marsh grass, barely above the level of the water, and they were simi-
larly built. Construction was of the slightest, the nest walls being
formed largely of growing grass, and the bottom of the same sort of
grass, some placed there but mostly consisting of the crushed vegetation
that had been growing on the hummock. The upper rim of the nest
wall was dry, but the lower two-thirds was sodden and the eggs were
wet. There was no down in any of the nests. In each case the parent
bird could slip off of her eggs directly into water deep enough to swim
in, six or eight inches in depth, and, by narrow channels through the
hummocks, she could reach an open pond nearby.

On July 7 many male scaups were seen that were assuming the
eclipse plumage. Two drakes noted July 18 were mostly in eclipse,
but were strong on the wing. On July 20 the first downy young
appeared. On September 19, the last date on which I visited a suitable
spot for this species, several small flocks were seen, perhaps twenty-five
or thirty birds, all told.

Four specimens were collected (nos. 44633-44636), one adult male,
one adult female, and two downy young.

Glaucionetta islandica (Gmelin). Barrow Golden-eye

Present at Carcross when we arrived, May 22. One of the most
abundant ducks about Atlin; nearly as numerous as the lesser scaup
and of even more general distribution. The difference in habitats of
the two species lay in the greater abundance of the golden-eye in
the little mountain lakes at high altitudes. We found no nests, but
the situation of some of the broods seen precluded the possibility of
their having been hatched within cavities in trees. Many families
were found on lakes above timber line, long distances from trees of
sufficient size to afford sheltering holes.

The first young appeared July 3. This particular brood was kept
under observation until August 13. Originally nine in number, it was
reduced (cause unknown) to seven by July 7. On August 13, the flock
consisted of six, plus the mother, and the young then were about three-