el 
Table 2. Fall observations of migrating Aleutian Canada 
geese in the Aleutian Islands, Alaska. 

No. of Source (personal 
Date Location geese communication) 
26 Aug 1965 Amchitka Island 40 RK. Wade 
4 Sept 1974 At sea east of 40-502 G. Putney 
Shemya Island 
5 Sept 1975 At sea 16 km east 17 ~——G. Putney 
of Buldir Island 
7 Sept 1976 — Kiska Island 12 ~=T. Dowell 
7 Sept 1976 Unalga Island 346 W. Hoffman 
(east) 
22 Sept 1976 Unimak Island gb J. Nelson 

“Probably birds released at Agattu Island in May 1974 (cf. 
Springer et al. 1978). 
bProbably B. c. leucopareia. 
cover. Since Buldir has no lakes in the lowland, timing of 
freshwater thaw was not important in nest site selection. 
In the lowland plant association, the beach rye-umbel 
and the beach rye-umbel—fern communities covered 68 % 
of the area. The sedge-fescue meadow community ac- 
counted for 24% of the association and was confined 
largely to the higher elevations of the Lowland Tall-plant 
association, often the transition area between the two asso- 
ciations. Less than 1-ha patches of sedge-fescue meadow 
were interspersed among the beach rye-umbel and beach 
rye-umbel-fern communities. 
Vegetation Selection for Nest Sites 
Geese nested almost exclusively in the beach rye-umbel 
and beach rye-umbel-fern communities. In both com- 
munities geese selected areas with dense concentrations 
of beach rye in the overstory and red fescue in the middle 
story (Table 3). The nest territories (5- x 5-m plots) had 
significantly higher densities of these plants than did the 
communities at large, and the nest sites (1- x 1-m plots) 
had significantly higher densities of beach rye, but not red 
fescue, than did the nest territories (Table 3). There may 
have been a slight selection against wild celery, and birds 
clearly avoided areas with high concentrations of cow 
parsnip at nest sites. Spring beauty (Claytonia sibirica) was 
also avoided. The percentage of fern in the overstory or moss 
in the ground story was not important in nest site selection. 
Beach rye must have been favored for its value as nest 
cover because dried hummocks provided protection from 
wind and rain. Green shoots usually appeared by the end 
of egg-laying (Fig. 9), and by mid to late incubation the 
new growth provided substantial cover. Beach rye was also 
the major plant used in nest construction, occurring in all 
nests. Based on observations of feeding geese and the abun- 
dance of cropped stems, red fescue was the major food taken 
by geese during incubation. It was not favored for nest 
material, however, and occurred in only 5% of the nests. 
Cow parsnip and wild celery developed later than beach 
rye, and their value as nesting cover was relatively low. 
Furthermore, cow parsnip collected considerable water in 
Table 3. Comparison of percent cover values of major plants at goose nests (1-m? and 25-m? plots) and in plant communities 
at Buldir Island, Alaska. 
Percent cover values? 
A—Nest 

Plant Story 
Beach rye-umbel n = 45 
Elymus arenarius over 65.6 + 2.9¢ 
Angelica lucida over 13.0 + 1.9 
Heracleum lanatum over 8.7 + 2.0 
Festuca rubra middle 32.4 + 3.0 
Claytonia sibirica middle 10.3 + 1.9 
Carex macrochaeta middle 9.5 + 2.6 
Moss ground 32.0 + 3.9 
Beach rye-umbel-fern n= 16 
Athyrium felix-fem over 33.7 + 5.5 
Elymus arenarius over 38.0 + 4.9 
Angelica lucida over 12.5 + 3.9 
Heracleum lanatum over 9.2 + 3.8 
Festuca rubra middle 30.3 + 7.7 
Claytonia sibirica middle 14.2 + 4.2 
Moss ground 33.6 + 7.9 
Significant differences» 
B—Territory C—Community A-B B-C A-C 
mn = 45 n= 143 
45.9 #£ 2.7 37.6 + 0.3 0.001 0.01 0.001 
15.8 + 1.0 16.5 & 0:2 ~— — 0.1 
20.7 + 2.1 20.9 + 0.3 0.001 0.001 0.001 
29.1 + 2.4 16.9 + 0.3 — 0.001 0.001 
14.8 + 2.2 22.3 + 0.3 — 0.01 0.001 
11.5.4 2.1 3.5 + 0.1 — 0.001 0.05 
so. & 34 36.1 + 0.3 — — — 
n= 16 n = 95 
27.1 + 4.3 24.4 + 0.5 — 0.1 — 
29.7 + 4.1 24.4 + 0.5 - 0.01 0.001 
12.6 + 1.5 12.9 + 0.2 — -- — 
23.7 + 2:5 20.6 + 0.4 0.01 _ 0.01 
928.4 + 4.4 3.4 + 0.2 — 0.001 0.01 
24.4 + 3.6 40.1 + 0.5 0.1 0.001 0.001 
31.0 + 5.0 40.4 + 0.6 — 0.1 — 

aA = percent cover value (PCY) at nest site (l- x 1-m plot); B = 
transect plots (1 x 1 m). 
PCV in territory plots (5 x 5m); and C = PCV in community 
bStudent’s “t-distribution” used to calculate highest critical probability at which a significant difference existed (— indicates no signifi- 
cant difference). 
‘Mean + standard error. 
dn = number of plots. 
