una of the Wakarusa River has
three major components:

(1) A group of species that are mainly restricted to the lower
mainstream; all of them are common in the Kansas River (_Lepisosteus
osseus_, _Carpiodes carpio carpio_, _Ictiobus cyprinella_, _Moxostoma
aureolum_, _Cyprinus carpio_, _Hybopsis storeriana_, _Hybopsis
aestivalis_, _Notropis percobromus_, _Hybognathus nuchalis_ and
_Pylodictis olivaris)_.

(2) A group of species that are ubiquitous; they comprised the entire
fauna in some tributaries, despite the existence of habitats that
seemed suitable for other species (_Notropis lutrensis_, _Pimephales
promelas_, _Ictalurus melas_, and _Lepomis cyanellus_).

(3) A group of species having distributions centered in Rock Creek,
Washington Creek, Deer Creek, and some nearby tributaries (_Catostomus
commersonnii_, _Semotilus atromaculatus_, _Hybopsis biguttata_,
_Notropis cornutus_, _Notropis topeka_, _Notropis umbratilis_,
_Phenacobius mirabilis_, _Pimephales notatus_, _Campostoma anomalum_,
_Noturus exilis_, _Percina caprodes_, _Etheostoma nigrum_ and
_Etheostoma spectabile_).

The distributions of groups (2) and (3) provide clues to the effect of
drought on the fish-population, and on the relative ability of various
species to repopulate areas where they have been extirpated.

Larimore _et al._ (1959) studied the re-establishment of stream-fish
following drought in Smiths Branch, a small warmwater stream in
Illinois. They found that 21 of the 29 species regularly occurring
there reinvaded most of the stream-course within two weeks after the
resumption of normal flow, and that all but three species were present
by the end of the first summer. Our study indicates a much slower rate
of dispersal by many of the same species. This is presumably
attributable to the ecological barrier presented by the Wakarusa
mainstream.

During the drought (1952-1956) the mainstream with its turbid water and
mud bottom could hardly have served as a refugium for species requiring
the clear water and gravel bottom of upland tributaries. Probably the
main refugia for these species [group (3)] were in the upper portions
of Rock Creek, Washington Creek and possibly Deer Creek. While
collecting we observed that these creeks had larger proportions of
gravel-rubble bottom, clearer water, deeper pools, and appeared to be
more stable than other creeks in the drainage. In Washington Creek,
Lone Star Lake enhanced stability of flow.

At the