52 
Rowland M. Shelley 
1989), although it is beginning to fragment as evidenced by the 
segregated populations of S. insulanus and bergrothi. Scytonotus is 
likewise cohesive in the east, particularly in the Blue Ridge Prov- 
ince, but the overall generic picture is so dominated by the three 
large lacunae and the three areas of the granulatus group (Fig. 33), 
as to create the impression of a highly dissected taxon. The bergrothi 
group, the only lineage west of the Cascades, is restricted to the 
Pacific Coastal region aside from the small area of S. insulanus in 
southeastern Washington; the inornatus group, essentially parapatric 
to the bergrothi group and S. simplex , and the allopatric granulatus 
group cover the other areas east of the Cascades. The absence of 
lacunae in the bergrothi group, in contrast to the granulatus group, 
suggests more recent evolution. The essentially parapatric occurrence 
of S. inornatus, the most plesiomorphic species, coupled with the 
nearly total dominance of the granulatus group to the east, point to 
the general area of the Cascades in southern Oregon as the primary 
source area or center of evolution. A secondary center exists in the 
Blue Ridge Province, and the distribution of S. granulatus appears 
to have been “unzipped” in a northeastwardly direction along the 
spine of the Blue Ridge beginning in north Georgia (Fig. 32). The 
greater distribution of S. australis and its spreading beyond the moun- 
tains onto the western periphery of the Piedmont Plateau suggest 
that it evolved first among the Blue Ridge endemics and has had 
time to expand and to attain reproductive isolation. In contrast, its 
younger counterparts to the north are still linked by intergrades and 
display progressively narrower ranges within the confines of this 
physiographic province. 
Relationships 
Generic — A cladistic analysis of genera in the Polydesmidae is 
beyond the scope of this work and is also impossible today, as 
many genus-group taxa are poorly known and consist of only one 
or two species. Many decades will doubtlessly pass before the al- 
pha taxonomy of the family is sufficiently stable to allow such an 
effort. Hoffman (1979) did not even attempt to divide the family 
into subfamilies, merely listing the component genera in alphabetical 
order and inadvertently omitting two monobasic ones: Mastodesmus 
Carl, 1911, from Java, as noted by Golovatch (1991), and Alpertia 
Loomis, 1972, from Washington state, United States. Consequently, 
only the most tenuous surmisals are possible about the affinities of 
Scytonotus. With two branches to the telopodite, it differs from the 
other native Nearactic polydesmid genera, so its relationships must 
lie with taxa on other continents, and Cook (1911) suggested, but 
