652 
THE WILSON JOURNAL OF ORNITHOLOGY . Vol 123. No. 3. September 2011 
wintering birds were assigned to a core sample on 
the basis of phenotypic similarity to banded birds. 
Within-taxon analysis of subsamples of the core 
taxa (the plumage characters mentioned earlier) 
showed that wintering birds were not distinguish¬ 
able from the known breeding samples—an 
unsurprising completion of the circle. The core 
taxa are distinguishable from most of the four 
surrounding taxa. The conclusions of this section 
seem to provide a mixed message. On page 101, 
there are cautions against recognizing taxa on the 
basis of small samples of wintering specimens. On 
page 103, “Use of wintering specimens to make 
taxonomic decisions is as accurate as when using 
birds from the breeding grounds.’’ Nonetheless, 
Anderson concedes (page 105) that “Hanson has 
named perhaps 4% too many subspecies where 
samples of 20 or more are involved." There 
seems to be no conceptual distinction made 
between assigning winter specimens to a named 
subspecies, and defining and naming a subspecies 
on the basis ol wintering specimens. 
A major conclusion of this analysis is that the 
AOU subspecies moffitti and canadensis are 
geographically variable and, thus, are not taxo- 
nomically valid as currently construed. Curiously 
the author wonders (page 106) “where did the 
original taxonomists who named them go 
wrong?” It is true that Linnaeus had a small 
sample (introduced birds) when he named the 
species, eventually subspecies, canadensis and 
probably did not recognize the extent of variation, 
but Anderson’s statement suggests a total lack of 
understanding of how the entire pre-Hanson 
subspecific treatment of geese developed. 
Table 6.1 is a comparison of Hanson’s taxa and 
those of the AOU—a welcome approach to a 
synonymy of Hanson’s many names. For each 
AOU subspecific name there is an alphabetical list 
of Hanson’s included taxa with a “potential” 
natal area and number of specimens. What the 
latter refers to is neither explained nor obvious* 
the number is frequently •zero', and occasionally 
? . The AOU race occidentalis is included in 
fidva, although the former has priority if the two 
are lumped. The name taverneri is in the AOU list 
although the AOU did not recognize that subspe¬ 
cies. Included with Hanson’s subspecies are 
several names not used by Hanson but that are 
used, although never formally defined or pro¬ 
posed, m this volume. There is some confusion in 
listing some Hanson names in cither parvipes or 
mterwr. There is a series of 20 Hanson names 
whose association with AOU names is ‘uncertain’ 
and for several others it is *?.’ Again, Vancouver 
ensis is incorrectly used as an AOU name. The 
AOU subspecies leucopareiu is elevated to 
species rank (as did Hanson) but minima is either 
ignored or merged with hutchinsii. This is an 
unsatisfactory synonymy in many respects. The 
rest of Chapter 6 is devoted to pairwise compar¬ 
isons of the amount of variation in the AOU 
subspecies, showing that the forms minima, fulva. 
and occidentalis are diagnosable from one another 
and the rest of the taxa but the rest are not 
separable from one another because of the great 
amount of variation in each. 
Analyses of segments of AOU populations in 
Chapter 7 show the individual segments (actually 
Hanson subspecies) are distinguishable from one 
another even though the larger populations are 
not. This seems to be essentially a repeat of the 
analyses ot Chapter 5. Chapter 8 shows that 
variation in the large forms of W-c G does not fit 
a pattern ol clinal variation, contrary to statements 
and assumption of many past workers. In small 
geese, there is a trend of increasing size with 
increasing longitude east of Alaska (Chapter 9). 
Chapters 10 presents a cluster analysis ol' large 
W-c G, which results in a dendrogram purporting 
to show the relationships of 71 of Hanson’s taxa 
plus lour recognized but not formally named by 
Anderson, including 2,555 individuals (also said, 
and counted, to be 74 taxa). The data are from 
pairwise comparisons using discriminant function 
analysis Irom earlier in the book. The dendrogram 
is interpreted as consisting of nine major stems, 
each with one or more minor stems. Each major 
stem (or clade) includes 3-16 phenotypically 
similar taxa, which may be widely separated 
geographically and associated with one of seven 
AOU subspecific taxa or tw r o of Hanson s species 
level taxa. Chapter 11 does the same kind ot 
analysis with 1.712 individuals in 54 taxa (55 by 
my count) of the small W-c G, resulting in six 
stems of Hanson taxa in five AOU subspecies cir 
three ol Hanson’s species. As summarized, this 
analysis groups taxa described by Hanson on the 
basis of phenotypic characters with clusters of 
similar taxa having wide, disjunct distributions; 
these similar taxa arc separated by phenotypically 
different taxa in other clusters, in something of a 
patchwork pattern. Anderson assumes (page 233 1 
this phenotypic similarity indicates evolutionary 
relatedness.’ Both the word and the concept 
convergence ’ seem to be absent from the book. 
