HARVARD UNIVERSITY

Library of the

Museum of

Comparative Zoology

MCZ
LIBRARY

OCCASIONAL PAPERS ^pp q q 199]

of the ^^^

MUSEUM OF NATURAL >tti§TbR^/
The University of Kansas
Lawrence, Kansas

NUMBER 140, PAGES 1-17 21 MARCH 1991

SYSTEMATICS AND BIOGEOGRAPHY OF NORTH
PACIFIC SHAGS, WITH A DESCRIPTION OF A

NEW SPECIES

Douglas Siegel-Causey'

The most derived clade of shags (Phalacrocoracidae: Leucocarboninae)
are the CUffShags {Stictocarbo spp.), a diverse assemblage of slender-bodied
marine birds restricted to neritic waters and coastlines. Cliff Shags are found
in New Zealand waters, Europe, the west coast of the Americas, and the north
rim ofthe Pacific Ocean (Siegel-Causey, 1988; Siegel-Causey and Litvinenko,
1992). Two species are known from North America: Pelagic and Red-faced
Shags (Stictocarbo pelagicus and S. wile). These birds are widely distributed
in the North Pacific basin and are sympatric throughout Beringia and the
North Pacific, including the Aleutian and Kurile Islands, the Soviet Far East,
and Sea of Japan.

Recent archaeological excavation of Aleut kitchen middens on Amchitka
Island in the far western part ofthe Aleutian Islands provided an opportunity
to examine the population history of Aleutian shags. My study of this material
(Siegel-Causey and Lefevre, 1991) revealed a diverse fauna of cormorants
and shags, some of which are not resident on Amchitka Island at present.

Throughout all strata and sites of these middens, I found in total 109
elements of a small phalacrocoracid that were not referable to any known
species. The majority were bones ofthe limbs and central core, although there
were a few cranial and mandibular fragments, furculae, and a sternum. I found
diagnostic qualitative characters on the mandible, furcula, humerus,

' Museum of Natural History, Dyche Hall, and Department of Systematics and
Ecology, Haworth Hall, The University of Kansas, Lawrence, Kansas 66045-2454,
USA.

2 UNIV. KANSAS MUS. NAT. HIST. OCC. PAP., No. 140

carpometacarpus, synsacrum, femur, tibiotarsus, and tarsometatarsus that
served to refer these elements to the genus Stictocarbo and to discriminate
them from Pelagic and Red-faced Shag bones (Appendix 1 ). Furthermore, all
of these bones were mensurally smaller than any of the extant species on
Amchitka Island (Table 1 ), and could be sorted easily from the other midden
elements by size alone.

Size is unreliable as a character, however, when the scope of comparison
broadens beyond a single locality. Pelagic Shags, for example, demonstrate
substantial clinal variation in size over their distribution (Siegel-Causey, in
prep.) and some specimens collected from the southern limits of the range
approach the size of these enigmatic midden bones. For these and other
reasons, I will employ robust analyses of the qualitative and quantitative
patterns of morphometric variation in these North Pacific shags rather than
rely on more simplistic univariate and descriptive accounts.

The discovery of a previously unknown shag on Amchitka Island prompted
me to search among all existing museum specimens from Beringia for
evidence of more recent presence in this region. Because external characters
are unknown and the only specimens are skeletons, it was not possible to
identify museum skins of this species. I found three skeletal specimens, all
collected coincidentally from Amchitka Island in the late 1950s, possessing
all of the qualitative characters identified from midden bones and within the
expected size range. This wealth of data makes possible a more detailed
analysis of the phylogenetic status and relationship of this new taxon.

METHODS

Subfossil material and skeletal specimens.— Bird bones were collected
from two sites on Amchitka Island by employees of Archaeological Research,
Inc. during the summer of 1969 and were later sorted to family by Stuart
Warter (see Acknowledgments). All subfossil material is kept at California
State University at Long Beach and was made available to me through the
auspices of Warter. Full details on the provenance of these bones and other
information relating to Amchitka Island are given in Siegel-Causey and
Lefevre (1991). I obtained from various sources (see Acknowledgments)
comparative skeletal specimens collected at or near breeding colonies in
Beringia; a list of specimens and museums is available from the author.
Museum specimens are identified by accession number and the following
acronyms: AMNH (American Museum of Natural History), CSULB
(California State University, Long Beach), KU (Museum of Natural History,
University of Kansas), LACM (Los Angeles County Museum of Natural
History), UMMZ (Museum of Zoology, University of Michigan), USNM
(U.S. National Museum of Natural History), UWBM (Burke Museum,
University of Washington), and UWZM (Zoological Museum, University of
Wisconsin).

SIEGEL-CAUSEY. 1991

NORTH PACIFIC SHAGS

Table 1 . Univariate statistics for skeletal characters of North Pacific Shags, genus
Sfictocarho. Measurement codes appear in parentheses in the character column and
refer to Figure 12 in Ono (1980). NA means not available.

Character

1.2

wile
(»= 17)

pelagicus

{n = 20)

kenyoni
(/; = 3)

midden bones
(/;= 109)-^

Cranium

L^(NA) 53.8

GW^'(NA) 30.8
frontonasal W^CNA) 12.8
postorbital W^NA) 24.5

Maxilla

L^NA) 59.1

GW^(NA) 11.0

D'-CNA) 5.9

Coracoid

L^(Ca) 70.3

sternal W-' (Cf) 22.3

Humerus

La (Ha) 134.1

shaft DWM He) 7.9

head W^' (NA) 24.4

distal WMNA) 15.0

Radius

L^^ (Rb) 144.4

shaft PW(NA) 5.0

distal W*^ (Re) 8.8

Ulna

LMUb) 143.3

shaft PW'-- (Ui) 13.1

distal WMUh) 10.8

Carpometacarpus

total L'^(Ca) 64.1

shaft PW (Cf) 4.7

distal W''(Cc) 13.5

Synsacrum

total Lb (Pc) 109.5

acetabular D (Pb) 18.6
acetabular W'^(Ph) 35.3

Femur

total LN Fa) 61.3

neckGWt'(Fe) 15.8

shaft DW^ (Fc) 7.6

distal W^iFh) 15.6

(0.5^^
(0.2
(0.3
(0.4

(0.5
(0.2
(0.2

(0.9
(0.3

(1.5
(0.1
(0.3
(0.2

(1.8
(0.4
(0.1

(1.7
(0.2
(0.1

(0.7
(0.1
(0.1

(1.1

(0.2
(0.4

(0.7
(0.3
(0.1
(0.2

51.8

(0.6)

28.5

(0.2)

12.6

(0.2)

22.1

(0.4)

56.2

(0.6)

10.8

(0.2)

22.1

(0.4)

65.3

(0.6)

20.3

(0.3)

125.7

(1.2)

7.3

(0.1)

22.5

(0.2)

13.6

(0.2)

129.5

(2.8)

4.5

(0.4)

8.4

(0.1)

133.1

(1.3)

11.5

(0.4)

10.4

(0.1)

60.6

(0.6)

4.4

(0.1)

12.7

(0.1)

103.4

(0.9)

16.9

(0.4)

33.0

(0.5)

55.5

(0.5)

14.2

(0.2)

7.1

(0.2)

14.4

(0.2)

46.3

0.9)

26.4

0.1)

9.9

0.7)

20.6

0.1)

50.9

0.2)

8.8

0.2)

4.1

,0.1)

57.7

0.2)

17.9

0.7)

108.4

2.5)

6.3

0.1)

20.0

0.2)

12.2

.0.4)

114.4

0.8)

4.2

1.5)

7.6 (

0.1)

118.9 (

;o.2)

10.2 (

0.2)

9.7 (

0.1)

54.4 (

0.1)

4.4 (

0.5)

11.5 (

0.2)

90.9 (

2.0)

16.3 (

0.5)

29.5 (

0.7)

49.1 (

1.0)

12.2 (

0.3)

6.7 (

0.2)

12.8 (

.0.1)

NA

NA

NA

NA

NA

NA

NA

56.7

(0.3) [10]

17.2

(0.9)

11.4

(2.7)[7]

6.7

(0.3)

20.4

(0.7)

12.3

(0.3)

13.3

(0.9) [4]

4.1

(0.8)

7.4

(0.4)

14.5

(0.3) [7]

10.0

(0.2)

9.5

(0.4)

54.6

(1.6)[19]

4.1

(0.4)

11.8

(0.4)

87.9

(2.5) [12]

16.2

(0.9)

31.2

(1.3)

49.6

(0.8)[15]

12.6

(0.4)

6.2

(0.4)

12.3

(0.4)

4 UNIV. KANSAS MUS. NAT. HIST. OCC. PAP., No. 140

Table 1. Continued.

urile

pelagicus

kenyoni

midden bones

Character'- 2

(« = 17)

{n = 20)

(n = 3)

{n = 109)3

Tibiotarsus

total L'' (Tib)

111.6 (1.1)

102.1 (0.9)

91.7 (1.1)

NA

head GW^' (Tif)

12.4 (0.1)

11.2 (0.1)

10.0 (0.2)

NA

shaft DWcTid)

6.9 (0.1)

6.6 (0.1)

5.4 (0.1)

NA

distal Wb(Tii)

12.0 (0.2)

11.1 (0.1)

10.0 (0.3)

NA

Tarsometatarsus

total Li't Tab)

55.6 (0.6)

53.0 (0.5)

47.8 (0.5)

47.4 (0.6) [14]

proximal GWb(Tac)

13.0 (0.2)

12.2 (0.1)

10.8 (0.3)

11.2 (0.6)

hypotarsus D (Taf)

15.8 (0.4)

15.3 (0.2)

13.9 (0.2)

13.8 (0.6)

hypotarsus W^ (Tae)

6.5 (0.1)

5.7 (0.1)

4.7 (0.1)

4.6 (0.4)

' D = depth, DW = distal width, GD = greatest depth, GW = greatest width. L = length,

LW = least width, PW = proximal width, W = width.
-Superscript letters indicate significance level of species differences in ANOVA: a =

P < 0.05, b = P < 0.001, c = P< 0.001.
'Elements are unassociated. Sample sizes are given separately in brackets.
^Mean(SEM).

Phylogenetic analysis. — Each of the qualitative osteological characters
used was a discrete trait in which I could obtain at least two discrete states. I
found several other characters having modal states in some specimens;
however, because the source of character variation in phalacrocoracids is little
known, I excluded these from the initial analysis. (See below for further
discussion on character variation.) I added them subsequently to the reduced
set and tested for stability of the resultant tree(s). If tree topology was
unchanged from the initial analysis, I used the character; if otherwise, I
excluded the character. In other words, I in effect mapped the variable
characters on the tree derived from the reduced data set and calculated
summary tree statistics from the augmented data set.

I determined polarities of each character using methods described in
Siegel-Causey (1988). Transformation series described as characters 13 and
14 were treated as unordered. The trees were described using PAUP 3.0G
(Swofford, 1990) and I used the Exhaustive Search method which guarantees
finding all possible trees. There were no differences in character placement
using the accelerated transformation (ACCTRAN ) or delayed transformation
(DELTRAN) optimizations. Complete details on terminology, methodology,
and rationale of phylogenetic systematics are given in Wiley, et al. ( 1 99 1). See
Appendix 1 for character descriptions and polarities and Appendix 2 for the
data matrix of character state codings for North Pacific shags.

SIEGEL-CAUSEY. 1991 NORTH PACIFIC SHAGS 5

Statistical analysis. — For mensural comparisons among taxa, I used only
specimens which had a full component of study elements: thus. I used partial
skeletons that had at least one of each element listed in Table 1 . but I excluded
all specimens with elements missing from this list. I did not measure sterna
because, for the sample of skeletons I was able to measure, more than half had
damaged or warped sterna.

The only exception to this policy involved use of the midden material in
univariate comparisons with the putative specimens of the new taxon.
Summary statistics of these measures are reported in Table 1, but no midden
bones were used in subsequent multivariate analyses.

Skeletal measures used in this study are a subset of those detailed in Ono
(1980); Table 1 references Ono's measurements where appropriate. Because
of the small number of specimens and the high number of unsexed specimens,
I pooled sexes even though cormorants and shags show great sexual
dimorphism. In all, I used40 specimens formultivariate comparisons (5. ////Ve:
10 males, 7 females; S.pelagicus: 4 males, 8 females, 8 unsexed; S. new species:
1 female, 2 unsexed).

I analyzed interspecific differences in skeletal measures using a stepwise
multivariate analysis of variance and ANOVAs of scores derived from
stepwise Canonical Analysis (CA) which were used to quantify maximum
interspecific differences. All of the multivariate techniques were performed
on log-transformed measurements. Statistical analyses were done using
computer programs available in BMDP (Dixon, 1988).

RESULTS

Description of new species. — The abundance of subfossil elements and
three extant skeletal specimens makes possible a description of this new taxon
of North Pacific shags.

Stictocarbo kenyoni, new species
Fig. 1

Holotype. — USNM 43 1 1 64, complete skeleton from an adult female from
Constantine Harbor, Amchitka Island, Alaska: collected 22 February 1 959 by
Karl W. Kenyon from a fish net. Mass of the specimen at the time of collection
was 2 lb 6 oz (1060 g).

Paratypes. — UWBM 186 13 and 186 14 (unsexed birds originally identified
as Phalacwcorax pelagicus, Constantine Harbor, Amchitka Island, Alaska,
collected 15 September 1957 by K. W. Kenyon). Subfossil skeletal material
(CSULB 1401 1-141 19) from Amchitka middens 31 and 36 is referred to this
species.

UNIV. KANSAS MUS. NAT. HIST. OCC. PAP.. No. 140

E

/^Sk^

D

^=^ "5

H

Fig. 1 . Type of Stictocorho kenyoni. new species (USNM 43 11 64). A. Skull and
mandible, left side. B. Humerus: top, anconal left side; bottom, palmar right side. C.
Carpometacarpus, anterior, left side. D. Coracoid, anterior, left side. E. Femur: top,
posterior left side; bottom, lateral right side. F. Synsacrum, posterior view. G.
Tibiotarsus, anterior left side. H.Ttarsometatarsus, posterior right side.

Diagnosis. — Stictocarho kenyoni is similar to S. wile but may be
distinguished from it SLwdS.pelagicus by its small size and by six autapomorphic
skeletal characters (Appendix 1 ). The fossa aditus of the mandible extends at
least Va of the length of the insertion of M. pseudotemporalis (character 1 ). The
attachment of M. dorsalis scapulae on the humerus is present as paired scars
in line on the bicipital crest (char. 5 ). The attachment of M. obturator externus
+ intemus on the feinur is deeply excavated and narrow (char. 13). The
attachment of M. flexor perforatus digiti II on the femur is deeply excavated
with a robust lateral bony crest (char. 14). The medial margin of the
attachment of M. flexor hallucis longis on the femur is coincident with the
medial prominence of the lateral condyle, causing its superior aspect to appear
sharply produced (char. 16). The supratendinal bridge of the tibiotarsus has
distinctly linear, parallel margins (char. 19).

Description. — The holotype shares two synapomorphic characters with
its congeners S. pelagiciis and S. mile. On the femur, the attachment of M.

SffiGEL-CAUSEY. 1991 NORTH PACIFIC SHAGS 7

flexor perforatus digiti III lies on the medial edge of the bone (char. 15), and
on the tarsometarsLis the medioplantar process of trochlea metatarsus II is
produced proximally (char. 24). It shares with S. wile seven unique skeletal
characters: (a) the anterior attachment of M. pectoralis superhcialis pars
superhcialis is parallel with the anterior edge of the furcula (char. 3): (b) the
ligamental furrow on the humerus is marked by a deep fossa on the medialmost
part (char. 6); (c) the infratrochlear fossa of the carpometacarpus is indistinct
(char. 9); (d) the area immediately adjacent to the supratrochlear fossa of the
carpometacarpus is unexcavated (char. 10); (e) the attachment of M. obturator
extemis on the synsacrum is unexcavated (char. 1 1 ); (f) the intercotylar fossa
of the tarsometatarsus is absent or convex (char. 22); and (g) only digit III has
a deeply excavated supratrochanteric dorsal fossa (char. 23). In addition, it
differs from S. pelagicus in that the attachment of M. adductus mandibulae
externus superhcialis on the mandible is indistinct rather than expressed as a
strong transverse ridge (char. 2), and also that the lateral edge of the humeral
pneumatic fossa is planar in the holotype rather than marked by a distinct
narrow excavation (char. 4).

External features. — External phenotypic features are unknown, except
for Kenyon's (pers. comm.) impression that the bills of the types were very
thin which is reflected by the significantly narrow width of the maxilla found
in the types (Table 1). 1 found several museum skin specimens originally
identified as Pelagic Shags that approached the small size of the type
skeletons. Because size is unreliable for identification in this clade (see
below), identification of these skins must remain problematic unless skeletal
elements contained within them provide positive means of identification.
Until additional specimens are collected or identified, the external appearance
of this species remains a mystery.

Description of paratypes. — Paratypes differ from the holotype only in
minor details related to size; skeletal material from the referred midden
material shows evidence of wear and aging.

Measurements. — See Table 1 .

Distribution. — Known only from Amchitka Island, Alaska.

Etymology. — This species is named after Karl W. Kenyon, who collected
all known extant specimens of this new shag, and whose pioneering work on
the Aleutian avifauna greatly advanced our understanding of the ecology of
North Pacific seabirds.

Common name. — Kenyon's Shag.

Skeletal specimens examined. — Stictocarho wile. USA: Amak Islands
(USNM, UMMZ, UWZM, 8 cf, 2 9), St. Matthew Island (USNM, 1 cT ), Pribilof
Islands (KUMNH, USNM, UMMZ, UWZM, 6 cT, 4 9), Southeast Alaska
(USNM, 5 cf, 3 9), Cape Thomson (KUMNH, 1 cf ); USSR: Kommandarski
Islands (USNM, Icf ); JAPAN: Hokkaido (USNM, 1 cX 1 9).

S. pelagicus. USA: West Coast (USNM, CAS, FM, 6 cf, 4 9, 1 ?), Southeast

8 UNIV. KANSAS MUS. NAT. HIST. OCC. PAP., No. 140

Alaska (USNM, UWZM, UMMZ, 4 cf, 3 9), Amchitka Island (USNM, UWBM,
5 cf, 2 9, 3 ?), Amak Islands, (USNM, UMMZ, 1 cf, 1 9), Pribilof Islands
(KUMNH, 1 cf, 1 9); CANADA: Mandarte Island (1 9); USSR: Kommandarski
Islands (USNM, 1 9), Siberia (USNM, 2 cf, 1 9); JAPAN: (USNM, 1 9).

MENSURAL COMPARISONS

Univariate comparisons. — Thirty-three of 37 skeletal measurements
differed significantly among species (Table 1 ); the other four were width or
depth measures which lacked significant power in discriminating among taxa.
Those measurements effecting the greatest difference among species were
length and greatest widths of limb elements and width of trunk elements and
cranium. Stictocarho kenyoni had the smallest means for all measures, S. wile
the largest.

None of the dimensions of the types differs significantly from means
obtained from individual midden elements (r-tests, P < 0.05). 1 conclude
therefore that the type specimen and unassociated midden elements represent
the same taxon. Midden elements were excluded from all multivariate
analyses.

Canonical analysis. — North Pacific shags were separated completely
(Fig. 2) by species using a stepwise canonical analysis of complete skeletons
(Wilk's X = 0.05 1 6, df = 1 9, 1 , 40). Two axes were sufficient to explain all of
the variance in canonical space. Prior identifications were by qualitative
characters; all pairwise discriminations between species were significant (f*
< 0.001, pairwise F-tests, df = 19, 22), and jackknifed classifications of
specimens to species were 100% correct. Nineteen of the skeletal variables
(Table 2) were incorporated significantly into the model (partial F-tests, P <
0.05). Species differences were significant on both canonical axes (ANOVA
on canonical scores, P < 0.0001).

Geographic variation. — The sample sizes were inadequate for all taxa to
allow a quantitative assessment of geographic variation in skeletons. It is
worth noting, however, that specimens collected from the southern limits of
the range of Pelagic Shags (Fig. 2, closed circles), while small and close in size
to Kenyon's Shag, were classified unambiguously by stepwise DFA as
Pelagic Shags. These small southern Pelagic Shags represent a lower bound
for multivariate variation in this taxa and none possessed the qualitative
characters found in Kenyon's Shag.

QUALITATIVE COMPARISONS

Phylogenetic relationships. — My preliminary observations indicated that
this new species is a shag (Leucocarboninae) and a member of the Pelagic-
Red-faced Shag clade (genus Stictocarho). Phylogenetic analysis using PAUP

SIEGEL-CAUSEY, 1991

NORTH PACIFIC SHAGS

1

■>^ 0
<

15 -1

u

o
c

U

0

urile

o
o

° 8

o
o

<h

o
o

-5

o °8

'OD

o pelagicus

"S^

0 5

Canonical Axis I

kenyoni

10

15

Fig. 2. Plot of scores of North Pacific Shags on canonical axes I and II. Specimens
were identified using qualitative osteological characters. The closed circles represent
Pelagic Shags collected from central California.

on 24 osteological characters and polarities given in Appendix 1 produced a
single tree. This tree (Fig. 3) has a length of 24 steps and CI = 0.923. excluding
autapomorphies.

Seven autapomorphies support the establishment of this taxon as a new
species. One (char. 5) relates to greater control of humeral rotation and
possibly more powerful flight (see Ou're, 1967). Three (char. 13, 14, 15) relate
to extension of the foot and flexion of digits and may represent increased
capabilities in underwater propulsion (see Owre, 1967). One (char. 7) is
homoplasious and is convergent to states observed elsewhere in the family
(e.g., Nesocarbo camphelli, Microcarho spp.) (Siegel-Causey, 1988).
Functions of the other characters are unknown.

Eight synapomorphies, six of which are unambiguous, establish the Red-
faced Shag as sister species to this new taxon. Two of the six (char. 3, 6) relate
to greater range in wing movement (see Owre, 1967), but the functions of the
remaining characters are obscure. Two (char. 12, 17) are homoplasious with
states present elsewhere in the genus and relate to more powerful action of the
hind limb (see Owre, 1967; Siegel-Causey, 1988). Two synapomorphies
(char. 15, 22) unite all three extant species as a clade; the first reflects a broad
degree of digital flexion, but the function of the other is obscure.

10 UNIV. KANSAS MUS. NAT. HIST. OCC. PAP.. No. 140

Table 2. Standardized canonical coefficients of skeletal variables on canonical
axes for North Pacific Shags, genus Stictocarbo (n = 40).

Standardized coefficient'
Character CA I CA II

BillGW 1.72 0.51

Coracoid L 4.37 -0.62

Coracoid sternal facet W 0.42 0.43

Humerus L -2.11 0.66

Humerus DW -1-09 -0.16

Radius L 2.24 0.10

Radius DW -0.68 -0.39

Carpometacarpus shaft W 0.88 -0.14

SynsacrumL 1-52 0.67

SynsacrumD -3.15 -0.66

SynsacrumW -2.16 -0.64

Femur L -1-07 -0.38

Femur head W -1-27 -0.08

Femur shaft DW -0.68 -0.28

Tibiotarsus shaft DW -1.30 0.74

Tibiotarsus DW 1.05 0.15

Tarsometatarsus L -1-61 0.14

Tarsometatarsus PD 0.73 -0.22

HypotarsusW -1-72 -0.44

Eigenvalue 29.9 1 .08

Percentage of variance 0.96 0.04

Among-species F - 31.04* 0.21

' Standardized by pooled within-group variances.

^ Significance levels of F-values (df = 2, 38): *P < 0.001

Specimen variation and identification. — An important factor associated
with skeletal specimens is the degree of confidence that can be given to field
identifications. In some cases, particularly in groups where species appear
similar or have been little-studied, misidentifications may be prevalent. There
seem to have been few mistakes in identification of most specimens of Pelagic
and Red-faced Shags collected as adults in breeding plumage. For all other
ages and plumages, however, identifications are often problematic and based
on typological constructs such as collection location or specimen size.
Mistaken identifications frequently occur because juvenile Red-faced and
Pelagic Shags are phenotypically similar and small Red-faced females and
large Pelagic males in basic plumage are confusing (Siegel-Causey, in prep.).
Therefore, it was critical to this study that 1 examined as many skeletons as

SIEGEL-CAUSEY, 1991 NORTH PACIFIC SHAGS 11

possible that were collected from known breeding localities. The number of
specimens encompassing the geographic range of the clade of North Pacific
shags is small; in all, I examined only 53 specimens not including the
specimens of Kenyon's Shag.

Using qualitative characters, I determined that eight of the skeletons I used
were misidentified (not including those subsequently identified to be Kenyon's
Shag). Four of these (UMMZ 224301, 224304, 224305, 224310) were
originally identified as Red-faced Shags, but closer examination of qualitative
characters revealed that they were probably Pelagic Shags, albeit large for
their gender. The other four (KU 60506, 60507, UWZM 15960, USNM
431371) were problematic and presented many difficulties in character
assignment. I coded these eight as unknowns, and excluded them from the
initial analysis. I discuss elsewhere (Siegel-Causey, in prep. ) the relevance of
these specimens in the context of widespread geographic variation.

For the most part, the remaining 40 specimens presented few problems in
identification or in assessment of character states. For all of the Pelagic Shags
used in this study, character states did not vary and I was unable to document
any variation other than minor differences associated with size or gender.
Nine characters, however, varied among the 17 skeletal specimens of Red-
faced Shags that I used, and these may indicate imprecise character state
definitions, intermediate states of mosaic characters, or natural variation
(Appendix 1 ). Within S. wile most of these state variants could be attributed
to underdeveloped or indistinct muscle scars or to pathology (e.g., UMMZ
224304 has oddly foreshortened humeri and concomitant change in the bony
evidence of muscle attachment). By contrast, two characters (12, 33) had
moderate variation among 10 specimens by character states intermediate to
those expected for any of the species. Deletion of these characters, however,
did not affect the stability of the tree shown in Fig. 3; further study on the
nature of character variation is needed before they are discarded from
consideration in phylogenetic analyses.

SYSTEMATICS AND BIOGEOGRAPHY

There is an additional member of the clade that includes S. pelagicus, S.
wile, and S. kenyoni. Chandler (1990) described a fossil shag from late
Miocene (Blancan, 2-3 mybp) sediments near San Diego, California. This
species, Stictocarbo kumeyaay, is represented by several upper and lower
limb elements having sufficient characters to allow him to establish it
unambiguously as belonging to the clade of North Pacific shags (see also
Appendix 1, Fig. 3). By close examination of these elements and characters,
I was able to determine that this fossil shag is primitive to the extant species.
Two characters were especially indicative of the derived state of the other
three shags. The cnemial crests of the tibiotarsus are spaced less than the width

12

UNIV. KANSAS MUS. NAT. HIST. OCC. PAP., No. 140

kenyoni

urile

20
18
14
13

a-b
a-b
a-b
a-b
a-b

pelagicus

4
2

Stictocarbo
(derived)

a-b
a-b

23

22

17

12

11

10

9

6

3

21
15

25

a-b
a-b
a-b
a-b
a-b
a-b
a-b
a-b
a-b

a-b
a-b

19

16

14

13

7

5

1

b
b
c
c
b
b
b

a-b

magellanicus

Stictocarbo

Fig. 3. Tree of the character changes for North Pacific Shags. Solid lines
represent derived characters and parallel lines are characters convergent elsewhere in
the family (not shown). The number to the left of the character state symbol is the
character number described in Appendix 1 ; the letter sequence to the right is the
character state transformation.

of the shaft in S. kumeyaay and outgroups, but in S. pelagicus, S. urile, and S.
kenyoni, they are wider than the shaft width. The attachment of M. flexor
peiforatus digiti III is on the internal edge of the femoral shaft in S. kumeyaay,
while in the others it is more medial (see Chandler, 1990, for more extensive
discussion). This makes S. kumeyaay a sister taxon to present day species of
North Pacific shags, but whether or not it is ancestral is problematic.

SIEGEL-CAUSEY, 1991 NORTH PACIFIC SHAGS 13

The union of S. kumeyaay and S. kenyoni with S. pelagicus and S. wile
implies a more complicated biogeographic history of this clade than was
expected before. Previously (Siegel-Causey, 1988), I proposed that Red-
faced and Pelagic Shags were descendents of an early branch of Cliff Shags
that left the Southern Ocean and moved northward along the west coast of the
Americas. The presence of 5. kumeyaay in the Blancan supports this notion,
and may also indicate that shags were recent arrivals to the west coast of North
America. All of the other fossil phalacrocoracids known from this region are
unambiguously cormorants (Phalacrocoracinae), and date from the early
Miocene to the Pleistocene (see Brodkorb, 1963; Olson, 1985). Until earlier
shag fossils are discovered, 5. kumeyaay represents the earliest date for this
clade in the North Pacific.

The pattern of relationships among North Pacific shags as shown in my
reconstruction of the phylogeny can be used to infer a plausible biogeographic
history of the group. The origins of this clade lie in the southern hemisphere;
at some time in the Pliocene an ancestral form dispersed from the Southern
Ocean into Meso- and North America. Given that most extant shags are
associated with cool temperate climates and cold neritic waters, this
transequatorial dispersal probably occurred during a major cooling event.
Subsequent hypsothermal elevation in temperature and change in
oceanographic conditions would have promoted vicariant speciation of the
common ancestor of the extant species. After this, it is difficult to speculate
on the likely biogeographic sequelae of evolutionary events.

The present-day distribution of extant North Pacific shags suggests that
successive vicariance events probably associated with Pliocene and Pleistocene
glaciations were the mechanisms for speciation rather than one of dispersal
and isolation. Pelagic Shags are amphiberingian in the broadest sense and
may circumscribe the ancestral distribution of this clade. The greatest
abundances of Red-faced Shags lie along the Kurile archipelago and
Kamchatkan coast; the Alaskan populations are relatively minor and dispersed
(Vyatkin, 1986; Siegel-Causey and Litvinenko, 1992). It is plausible that this
species is a Kurilian vicariant with later eastward dispersal into the lower zone
of the Bering Sea.

It is noteworthy that all of the misidentified shags were collected from the
Pribilof Islands and western Aleutian peninsula. I found a similar pattern of
a localized distribution of misidentifications using museum skin specimens
(Siegel-Causey, in prep.), which makes it unlikely that the misidentifications
were due to chance. The significance of these ambiguous specimens is still
unknown.

The biogeographic affinities of Kenyon's Shag are unknown. It is tempting
to speculate that this species represents an Okhotskian vicariant originating
through isolation in the early Pleistocene. Such a mechanism was proposed
for early radiation of Cepphus alcids into Beringia (Storer, 1952).

14 UNIV. KANSAS MUS. NAT. HIST. OCC. PAP., No. 140

It is frustrating that there are few clues to the external appearance of
Kenyon's Shag. The fact that there are no previous records of a distinctly
plumaged small shag in the North Pacific suggests that Kenyon's Shags are
phenotypically very similar to Pelagic Shags. There are many examples of
cryptic taxa within species groups in shags; for example, Antarctic and
Imperial Shags (Notocarbo bransfieldensis and^V. athceps; see Siegel-Causey
and Lefevre, 1989). and the confusing complex of New Zealand shags
{Nesocarho campheUi and Eideucocarho spp.; Voisin. 1973; Siegel-Causey,
1988 ). Moreover, Pelagic and Red-faced Shags are difficult to discriminate at
a distance in field conditions, and sometimes even close at hand (Siegel-
Causey, in prep.). Identifications based on locality or size further complicate
matters and obscure the nature of distributional patterns.

So far, Kenyon's Shag is known only from Amchitka Island, but I
conjecture that more individuals will be found westward, in Okhotskian and
Kamchatkan waters. If the collection dates of the types are indicative, then
they probably were winter dispersers from less-studied regions to the west.
Okhotskian cormorants and shags move east or south in winter (Siegel-
Causey and Litvinenko, 1 992), and there is historical evidence for other Asian
phalacrocoracids in the Aleutians (Siegel-Causey and Lefevre, 199 1 ). Further
details of the distribution and ecology of this new species await discovery.

ACKNOWLEDGMENTS

I thank Stuart Warter (California State University, Long Beach) for access
to his extensive collection of midden bones, and P. S. Humphrey and B. C.
Livezey for insightful comments. I thank the following curators and museums
for assistance in borrowing or examining specimens: L. P. Baptista (California
Academy of Sciences), G. F. Barrowclough (American Museum of Natural
History), M. C. McKitrick (Museum of Zoology, University of Michigan), S.
Rohwer (Burke Museum, University of Washington), E. E. Pillaert (Zoological
Museum, University of Wisconsin), and R. L. Zusi (U.S. National Museum
of Natural History). This research was partially supported by the Museum of
Natural History, The University of Kansas, Lawrence.

LITERATURE CITED

Brodkorb, p. 1963. Catalogue of fossil birds. Part 1 . Archaeopterygiformes through

Ardeiformes. Bull. Fla. State Mus. Biol. Sci. 7:179-253.
Chandler, R. M. 1 990. Recent advances in the study of Neogene fossil birds, II. Fossil

birds of the San Diegan formations, late Pliocene, Blancan. San Diego County,

California. Ornithol. Monogr. 44:77-161.
Dixon, W. J. (ed.). 1988. BMDP Statistical programs. UCLA, Los Angeles. 1234 pp.
Olson, S. L. 1985. The fossil record of birds. Avian Biol. 8:80-238.
Ono, K. 1980. Comparative osteology of three species of Japanese cormorants of the

SIEGEL-CAUSEY. 1991 NORTH PACIFIC SHAGS 15

genus Phalacrocora.\{Aves, Pelecaniformes). Bull. Nat. Sci. Mus., Sen C. 6(4): 129-

151.
OwRE, O. T. 1 967. Adaptations for locomotion and feeding in the Anhinga and Double-
crested Cormorant. Ornithol. Monogr. No. 6. Amer. Ornithol. Union, Washington,

D.C. 106 pp.
Siegel-Causey, D. 1988. Phylogeny of the Phalacrocoracidae. Condor 90:885-905.
Siegel-Causey, D., and C. Lefevre. 1989. Holocene records of the Antarctic Shag

(Phalacroconi.x [Notocarbo] hninsfieldensis) in Fuegian waters. Condor 9 1 :408-

415.
Siegel-Causey, D., and C. Lefevre. 1991. Historical biodiversity of connorants and

shags from Amchitka Island, Alaska. Condor. In press.
Siegel-Causey, D., and N. M. Litvinenko. 1 992. Cormorants and shags of the temperate

North Pacific, Bering, and Okhotsk seas. In: Vemieer, K., K. L. Briggs, and D.

Siegel-Causey (eds.), Seabirds of the temperate North Pacific. Canadian Wildlife

Serv., Ottawa. In press.
Storer, R. W. 1952. A comparison of variation, behavior, and evolution in the seabird

genus Uria and Cepphus. Univ. Calif. Publ. Zool. 52:121-222.
Swofford, D. 1990. PAUP for the Macintosh, Version 3.0G. Illinois Natural History

Survey.
VoisiN, J. F. 1973. Notes on the Blue-eyed Shags (genus Leucocarho Bonaparte).

Notomis 20:262-271.
Vyatkin, p. S. 1986. Nesting records of colonial birds of the Kamchatkan peninsula,

pp. 20-36. In: Litvinenko, N. M. (ed). Morskoi Ptitsi Dal'nego Vostoka. Akad.

Nauk SSSR, Vladivostok.
Wiley, E. O., D. Siegel-Causey, D. L. Brooks, and V. A. Funk. 1991. The Compleat

Cladist. Univ. Kansas Mus. Nat. Hist. Spec. Pub. In press.

APPENDIX 1

Character Descriptions

Mandible

1 . Fossa aditus: (a) extends usually '/* (but no more than '/:) the length of the insertion
of M. pseudotemporalis; (b) extends at least Y-i the length of the insertion.

2. Attachment ofM. adductits mandihidae externus superficialis: (a) absent, reduced,
or indistinct; (b) expressed as strong transverse ridge. (Char. 42 of Siegel-Causey,
1988.)

Furcula

3 . Attachment ofM. pectoralis superficialis par superficialis: (a) the anterior margin
is directed posterially near the acrocoracoid facet; (b) the anterior margin runs
parallel with the anterior edge of the furcula.

Humerus

4. Pneumatic fossa: (a) lateral surface immediately adjacent to bicipital crest is
subplanar; (b) lateral surface marked by distinct, narrow excavation.

16 UNIV. KANSAS MUS. NAT. HIST. OCC. PAP., No. 140

5. Attachment ofM. dorsalis scapulae: (a) proximalmost scar lateral to distal scar; (b)
both scars in line on bicipital crest.

6. Li gamental furrow: (a) entire length of equal depth; (b) medialmost part marked
by deep fossa.

7. Li gamental furrow: (a) does not reach head; (b) distinctly notches head.(Char. 64
of Siegel-Causey, 1988.)

8. Deltoid shaft: (a) laterodistal surface concave; (b) laterodistal surface strongly
convex. (Char. 69 of Siegel-Causey, 1988.)

Carpometacarpus

9. Infratrochlear fossa: (a) deeply excavated; (b) indistinct or obsolete.

10. Carpal trochlea: (a) area immediately adjacent to supratrochlear fossa broadly
and deeply notched into triangular depression; (b) area unexcavated or with a
narrow, shallow depression.

Synsacrum

1 1 . Attachment of M. obturator externus: (a) by distinct fossa; (b) indistinct or obsolete.

12. Attachment of M. iliacuspreacetabulae: (a) intramuscular line arises on posterior
edge of preacetabulum; (b) arises on lateral surface of preacetabulum. (Char. 110
of Siegel-Causey, 1988.)

Femur

13. Attachment ofM. obturator externus + internus: (a) elliptical, shallow, indistinct;
(b) deeply excavated, broad, subcircular; (c) deeply excavated, narrow. (Char. 5 of
Siegel-Causey and Lefevre, 1989.)

\A. Attachment of M. flexor perforatus digiti U: (a) indistinct; (b) deeply excavated
without noticeable lateral bony margins; (c) deeply excavated with robust lateral
bony crest.

15. Attachment of M . flexor perforatus digiti III: (a) on internal edge of shaft; (b) on

medial edge.

16. Attachment ofM. flexor hallucis longis: (a) medial margin marked by distinct line
just adjacent to medial prominence of external condyle; (b) medial margin
coincident with medial prominence, causing the superior aspect to appear sharply
produced.

17. Trochanter: (a) anterior angle distinct; (b) anterior angle indistinct, curvilinear.
(Char. 128 of Siegel-Causey, 1988.)

18. Intramuscular line between M. flexor perforatus digiti III + IV: (a) reduced or
indistinct; (b) produced into rugose crest. (Char. 120 of Siegel-Causey, 1988.)

TiBlOTARSUS

19. Supratendinal bridge: (a) lateral (inferior) width greater than medial (superior);
(b) widths equal.

20. Plantaris fossa: (a) shallow; (b) excavated.(Char. 131 of Siegel-Causey, 1988.)

21. Cnemial crests: (a) robust, close set; (b) thin, widely flared.

SIEGEL-CAUSEY, 1991

NORTH PACIFIC SHAGS

17

Tarsometatarsus

22. Intcrcotylar fossa: (a) deeply excavated with strong medial crest; (b) absent or area
convex.

23. Siipratrochanteric dorsal fossae: (a) excavated at digits II, III, and sometimes IV;
(b) excavated at digit III only.

24. Trochlea metatarsus II: (a) plantar curve simple; (b) strong medioplantar process
extends proximally. (Char. 134 of Siegel-Causey, 1988.)

APPENDIX 2

Data matrix of character state codings for North Pacific Shags, genus Stictocarho.
Codings for the hypothetical ancestor (not listed) are all primitive. Numbers and
codings refer to transformation series and characters described in Appendix 1 ; codings
are; 0 = a, 1 = b, 2 = c, m = missing, U = unordered.

Characters

ikumeyaay

pelagicus

wile

kenyoni

1

m

0

0

1

2

m

1

0

0

3

m

0

1

4

0

1

0

0

5

0

0

0

1

6

0

0

1

7

0

0

0

1

8

0

0

0

9

0

0

1

10

0

0

0

11

0

0

1

12

0

0

1

13 [U]

0

0

2

14 [U]

0

0

2

15

0

1

1

16

0

0

0

1

17

0

0

1

18

0

0

0

19

0

0

0

1

20

0

0

0

21

0

1

1

22

0

0

1

23

0

0

1

24

1

1

1

UNIVERSITY OF KANSAS

MUSEUM OF NATURAL HISTORY

PUBLICATIONS

The University of Kansas Publications, Museum of Natural History,
beginning with volume 1 in 1946, was discontinued with volume 20 in
1971. Shorter research papers formerly published in the above series are
now published as The University of Kansas Museum of Natural History
Occasional Papers. The University of Kansas Museum of Natural
History Miscellaneous Publications began with number 1 in 1946.
Longer research papers are published in that series. Monographs of the
Museum of Natural History were initiated in 1970. Authors should
contact the managing editor regarding style and submission procedures
before manuscript submission. All manuscripts are subjected to critical
review by intra- and extramural specialists; final acceptance is at the
discretion of the Director.

This publication is printed on acid-free paper. Occasional Papers and
Miscellaneous Publications are typeset using Microsoft® Word and Aldus
PageMaker- on a Macintosh computer. - Museum of Natural History,
The University of Kansas, Lawrence.

Institutional libraries interested in exchanging publications may obtain
the Occasional Papers and Miscellaneous Publications by addressing the
Exchange Librarian, The University of Kansas Library, Lawrence,
Kansas 66045-2800, USA. Individuals may purchase separate numbers
from the Office of Publications, Museum of Natural History, The
University of Kansas, Lawrence, Kansas 66045-2454, USA.

Editor. Robert D. Holt

Managing Editor: Joseph T. Collins

Design and Typesetting: Kate Shaw and Joseph T. Collins

Printed by

University of Kansas Printing Service

Lawrence, Kansas

DATE DUE

4ilAR 34 1996

DEMCO, INC 38-2931

3 2044 093 361 707