Although extremely polymorphic, the bulk of the North American material is not separable from much of the Eurasian material labelled T. montanum and T. cochleariforme and seems to fit well the description of the type of T montanum as recorded for me by Dr. Cronquist. There appears to be no morphological basis for taxonomic distinction between T. montanum, T. cochleariforme, and North American plants.
The one cross between var monlanum (Europe) and var montanum (U. S.) had 75% successful silicle set, and the Fi hybrids were fertile (as judged by 58-93% stainable pollen).
Future study will probably reveal recognizable varieties within T. montanum in Eurasia as there are in the United States, and eventually perhaps a combination of characters can be found to separate our material infraspecifically from the Eurasian.
A polygonal graph (Fig. 56) depicts the similarity of a sample of U. S. specimens to a sample of Eurasian ones.
Twelve controlled cross hybridizations were made between T. parviflorum and
T. montanum, resulting in good silicles 0-53% of the time. Pollen stainability in the
F1s ranged from 0-42%. T. parviflorum differs from T. montanum in its smaller,
narrower erect petals, smaller sepals, shorter styles, smaller, more numerous seeds
per silicle, narrower silicles, non-caespitose habit, self-pollinating breeding system,
and non-odoriferous flowers. By virtue of partial genetic barriers and morphological
dissimilarities (Fig. 57), T. parviflorum is considered to be specifically distinct from
The remainder of the taxa in the United States form a very polymorphic complex consisting of none-too-clearly delimited, but geographically distributed, elements which can be treated arbitrarily as species, or I believe more realistically as varieties of one Eurasian-American species. These taxa are recognized as varieties rather than subspecies by virtue of their restricted and more localized distribution.
Within T. montanum crossability ranges from 8-100%) and pollen stainability is high (34-100%) with the exception of low stainability from the crosses between diploids and tetraploids. Wide-ranging in western North America, the species consists of one widespread variety and four satellite varieties mostly scattered around the margin of the overall range, one in southwestern Oregon, others respectively in western California, Arizona, New Mexico and Texas, and in the mountains of central Idaho. These outlying varieties are much less variable than the widespread var montanum. Other patterns of variation, though evident in individual populations, will not correlate with geographic distribution, and since the ecological differentiation is so subtle, other taxonomic entities are not distinguishable.
By growth of four of the five varieties in the greenhouse it was possible to study which of the various features were genetically determined rather than environmentally induced. The diagnostic characters which I have chosen are those least influenced by the environment.
Riley (1956) had found T. alpestre of Europe to be predominantly self-pollinating. Since our plants often were considered to be synonymous with T. alpestre, it had generally been assumed that they were also self-pollinating. Field observations of insect visitations and greenhouse observations on self-pollination and self-compatibility suggest that T. montanum is primarily outbreeding. The clinal nature of morphological variation in the species can partially be explained on this premise. On the other hand, many of the populations seem to be well separated spatially from one another, so are undoubtedly inbred, regardless of the general outbreeding nature of the species.