PROC. N. A. S. 39: 74-79 (1953)
Communicated by P. A. Shaffer; read before the Academy, November 12, 1952

1 Woodson, R. E., Jr., Ann. Missouri Bot. Gard., 34, 353-432 (1947).

Asclepias tuberosa, commonly known as butterflyweed and orange milkweed, consists of four subspecies separable primarily on leaf shape. The subspecies are isoploid and cross pollinated, and hybridize introgressively at the juncture of their ranges.1

2 Asclepias tuberosa ssp. terminalis Woodson, subspec. nov.— Subspecei interiori affinis et cum ea vulgo conjugalis, foliis lanceolatis basi saepius cuneatis vel truneatis rarius minus cordatis inflorescentia minus floribunda differt. Typus: L. S. Ehlers 13954 in herb.— Missouri Botanical Garden (Burt Lake, Cheboygan Co., Michigan).

Of particular interest are subspecies interior and terminalis2 which are sympatric and occupy a vast area from the Great Lakes to northern Mexico. Interior bears leaves which typically are broadly cordate or lobed at the base, while the leaves of terminalis are cuneate or tapered at the base. Although one occasionally may find plants with leaves characteristic of one subspecies well within the range of the other, the most deeply cordate leaves are found in the greatest concentration in the central Ozarks of southeastern Missouri. Outward from this center leaves become on the whole less cordate and more cuneate progressively to the periphery of the species range on all sides.

This peripheral effect is known in other plants and animals. Study of its nature in Asclepias tuberosa is facilitated by a method of measuring the shape of the leaf base which I devised several years ago.1 This method, which is remarkably sensitive and accurate, would assign an angular value (called L B) of more than 90° to a cordate leaf base, of less than 90° to a cuneate leaf base, and of exactly 90° to a leaf base exactly perpendicular to its longitudinal axis, which the descriptive taxonomist would call truncate.

In my first study of leaf variation in Asclepias tuberosa, after measuring several thousand leaves from as many herbarium specimens within the range of ssp. interior and terminalis, I demonstrated quantitatively the centrifugal decrease of L B by assigning the various L B values obtained to appropriate circles inscribed upon a map and concentric from an arbitrary center in the Missouri Ozarks. Lastly, to determine what ecological and latitudinal factors might be involved, the artificial populations from the concentric circles were divided to supply opposed series of concentric arcs from the point of origin in the Ozarks to the northeast and to the southwest, respectively. The two curves of L B values (lower panel of Fig. 1), which I have not attempted to smooth, were found to be nearly identical in spite of the widely differing climatic conditions encountered-and particularly so in view of the limited samples available.

I have had cultures of both subspecies under observation in my garden for several years and the genetic constancy of the phenotypes has been confirmed. Under cultivation it is impossible for me to distinguish plants of ssp. terminalis from such remote areas as central Minnesota and southwestern Texas, as it is in the herbarium, whatever subtile ecotypic responses may be involved.

That two easily distinguishable populations such as ssp. terminalis and interior should assume their strikingly regular concentric ranges independently is quite unthinkable. It is even more so to consider the central ssp. interior to be the spontaneous product of hybridization from the periphery, which has been suggested to me.

The most reasonable interpretation, in my opinion, is to view ssp. terminalis as an ancient, ecologically very tolerant population, and ssp. interior as a relatively recent genetic innovation within it probably dating much more recently than the retreat of the glaciers, to judge from the remarkably even phenocontours both to northeast and to southwest.

It long has seemed to me that the peculiar concentric distributions of ssp. interior and terminalis provide a study in natural selection in which the presumably older terminalis is being supplanted by the younger, more aggressive ssp. interior through introgressive hybridization and some selective superiority transmitted by the latter to its heterozygous progenies. During the past summer opportunity to continue the problem in this light was given me by grants from the BACHE FUND of the NATIONAL ACADEMY OF SCIENCES and the Permanent Science Fund of the American Academy of Arts and Sciences.

To this end, an automobile trip was taken from about the middle of Circle 1 in the central Ozarks (middle panel, Fig. 1) due north to the limits of Arc NE 5, about one hundred miles north of Minneapolis. Thirty colonies of Asclepias tuberosa were encountered, and a random flowering stem taken from each plant. The colonies averaged about 25 plants. Although in previous studies only leaves were measured, in this case several additional measurements and counts were made.

In sum, it was found that the total number of flowers produced by a plant bearing cordate leaves (i.e., ssp. interior) usually is significantly greater than the number produced by plants with cuneate leaves (i.e., ssp. terminalis). Flowers in Asclepias are borne in small individual clusters called cymes, and do not open for pollinating agents all at one time, but progressively. The number of flowers borne in the cymes does not differ significantly in these two subspecies of Asclepias tuberosa, but a typical stem of ssp. interior produces more cymes than does one of ssp. terminalis. This accounts for the differing totals of flowers produced, and consequently the number of cymes has been used as the unit of measure in this study. Furthermore, like the flowers, the cymes do not mature at one time, but in succession.

It thus follows that the larger number of cymes produced by a typical plant of ssp. interior not only produces more pollen but also over a correspondingly longer period than the smaller number of cymes characteristic of a plant of ssp. terminalis. One might expect that ssp. interior would be a more effective pollinator than ssp. terminalis not only in initial hybridization but exponentially so in back crosses. I believe it reasonable to assume that this character of ssp. interior, together incidentally with a probably greater seed production as well, is of very positive selective value, and that the character of the cordate leaves, among others, is merely associated with it in its conquest of the older ssp. terminalis.

3 Vavilov, N. I., Studies on the Origin of Cultivated Plants. Leningrad, 1926.

Space does not permit extensive report of data collected during the past summer; a more detailed account will be given elsewhere. However, very interesting effects are obtained by projecting the colonial means of the thirty colonies sampled in Missouri, Iowa, and Minnesota upon a grid for cyme number and L B (upper panel of Fig. 2). Particular attention should be called to the high variability of the Missouri colonies and the relatively low variability, particularly for cyme number, of the Minnesota colonies as would be expected from Vavilov's3 theory of the centers of origin of cultivated plant species if our hypothesis were correct. It will also be noticed that the recombination pattern assumes the properties of an exponential curve.


Upper panel: Asclepias tuberosa sspp. interior and terminalis contrasted.


Middle panel: method of studying internal differentiation in A. tuberosa.


Lower panel: profile of genoclines in sspp. interior and terminalis. Explanation in the text.


Upper panel: colonial means of thirty colonies of Asclepias tuberosa sspp. interior and terminalis.


Lower panel: individual plants of three colonies of the above. Explanation in the text.

In judging whether introgressive hybridization is the implement of natural selection in Asclepias tuberosa, the lower scatter diagram of figure 2 is helpful. Here the individual plants of three colonies are plotted for cyme number and L B. A number of interesting and suggestive properties are seen. But of immediate interest are the distinctive and relatively uncorrelated distributions of the colonies from Minnesota and Missouri and the greater variability and higher correlation of the Iowa colony, wholly indicative of heterozygosity for these characters, I believe. Although both characters apparently are multifactorial, notice the apparent segregation of the Iowa colony into blocks simulating a monofactorial F2.

In my opinion the pre-eminent evolutionary significance of introgressive hybridization in both plants and animals is as a vehicle for natural selection. Indeed, particularly if systemic mutations are dismissed, it might be considered the principal vehicle in sexually reproducing organisms. No other mechanism could be more natural nor more subtile for the statistical transformation of populations essential to the Darwinian theory. I consider the process of hybridization under selection pressure to be so fundamental to organic evolution that, as an inveterate descriptive taxonomist, I am giving it a name: sabinism. This will recall the violent fusion of populations which formed Rome, particularly the early military and genetic conquest of the Romans over their neighbors the Sabines, and familiar to most of us through the painting of Louis David.

  1. Woodson, R. E., Jr., Ann. Missouri Bot. Gard., 34, 353-432 (1947).
  2. Asclepias tuberosa ssp. terminalis Woodson, subspec. nov.— Subspecei interiori affinis et cum ea vulgo conjugalis, foliis lanceolatis basi saepius cuneatis vel truneatis rarius minus cordatis inflorescentia minus floribunda differt. Typus: L. S. Ehlers 13954 in herb.— Missouri Botanical Garden (Burt Lake, Cheboygan Co., Michigan).
  3. Vavilov, N. I., Studies on the Origin of Cultivated Plants. Leningrad, 1926.