The Journal of Heredity. 17: 413-426 (1926)
Metaphanic Variations in Rose Sepals
Inter-Organ Hybrids between Leaf and Calyx—
Significance of Such Variations in Study of Evolution and Heredity

O. F. Cook
Bureau of Plant Industry

THE study of heredity has revealed a world of new facts and relations that are impossible to describe without extending our vocabulary. The subject-matter of heredity is different from any that the mind of man has dealt with in the past. In these newly explored regions the paths of thought have not been opened or set with words in convenient places, to mark the distinctions that are necessary.

The word "metaphanic" was chosen to convey the sense of changed expression of characters, where characters appear outside their usual position or sequence of development, or in combinations with other characters that normally are distinct, but sometimes are confused or united in expression. Metaphanic variations are those that show modified or intermediate expressions of characters. Leaf and calyx, for example, are normally distinct and contrasted in their characters, but when intermediate organs occur, as sepals showing characters of leaves, they may be described as metaphanic variations.

Many variations may be considered as metaphanic, though only a few have been recognized and studied from the standpoint of the expression relations of the characters, as distinguished from other types or classes of variations. No word that is shorter or more familiar than metaphanic has been found to convey the idea of modified expression of characters.

Many of the metaphanic variations are plainly abnormal, as when abortive organs are formed which are intermediate between petals and stamens, or between carpels and leaves, but other metaphanic variations may not interfere with reproduction, as in the present examples of intermediates between the foliage leaves of roses and the sepals.

*Brachysm, a Hereditary Deformity of Cotton and Other Plants, Journal of Agricultural Research, 3:387, Feb 15, 1915.
   Diversity of Internode Individuals, Journal of Heredity, 13:323, July, 1922.
   Evolution of Compound Leaves in Walnuts and Hickories, Journal of Heredity, 14:77, May, 1923.
Several examples of metaphanic variation in cotton and other plants have been noted in previous publications.* The cotton plant shows many intermediate stages between the foliage leaves and the involucral bracts and also stages between the involucral bracts and divisions of the calyx.

The illustrations of rose sepals now offered may render the subject of metaphanic variation more intelligible. The material was obtained in December, 1924, at Shafter, California, and some of the most striking cases are shown in the natural-size photographs reproduced in Figures 10 and 11.

Under the late-season conditions several varieties of roses showed a tendency to foliar enlargement of the calyx, in forms intermediate between foliage leaves and normal sepals. The variations might be described as phyllody of the calyx, though the abnormalities seem less than in the so-called "green roses" where the petals are replaced by more or less leaf-like organs. Unfortunately, notes were not kept of the varieties from which the photographs were taken, but similar variations occurred in several standard varieties, including American Beauty, which is believed to have contributed most of the specimens that were studied.

Considered as products of abnormal development, these metaphanic sepals may be ascribed to delay or postponement of the changes of characters that occur between leaves and sepals. Reductions in the size of the leaves that occur normally in advance of the formation of the calyx are deferred to the stage of formation of the sepals, so that characters of leaves are shown in different degrees in the modified sepals. In the normal calyx the sepals are much smaller and more nearly equal, and the uppermost leaf below the calyx in some varieties is reduced to a simple bract, often quite similar to the sepals. A series of such bracts, reduced to the size and nearly to the form of the normal stipules is shown at the top of Figure 13.

Leaves that appear in the next position below a normal calyx usually are much more reduced and simplified than the leaves shown in Figure 11. In that position the leaves should have been much more reduced, and in form more like the metaphanic sepals. Thus the leaves in Figure 11, while not abnormal in form, yet were abnormal in their positions at the next node below the calyx.

Hybrids Between Metamers

Metaphanic variations may be considered as hybrids between different kinds of metamers of which the normal plant body is composed. Each plant may be analyzed into a series of different types of internode individuals, or metamers, as formed in the successive stages of the normal life history. The types of metamers show many contrasting characters, and the transitions from one type to another may be very abrupt. The first metamers of plants, represented by the so called cotyledons or seed-leaves, usually are very different from the next following metamers. Also the floral envelopes are different from the foliage leaves, as well as from the essential organs of the flower.

In rare cases metamers with intermediate characters occur, instead of the normally contrasted metamers, and these are reckoned as metaphanic variations. Some of them could be described very well as interorgan hybrids, if the word interorgan were not so easily confused with interorganic, thus leading to ambiguity. The word interorganic would suggest the idea of combinations between different organisms, which would be very misleading.

Metamerism in Plants and Animals

The principle of metaphany appears more prominently in plants than in animals on account of the nature of the plant body, which is built of a succession of structural units, the internode individuals, or metamers. While the bodies of animals are also segmented or metamerous, the methods of development are not the same as in plants. In all of the higher types of animal life the metamers are formed in the early embronic stages, and the whole body develops together, instead of increasing by a gradual addition of newly-formed metamers. Thus a simultaneous metamerism in animals may be contrasted with a successive metamerism of plants.

From the nature of the plant body it is plain that the same radical distinctions are not to be maintained between somatic tissues and germ-plasm as have been inferred from the study of animals. Or if the distinctions are assumed theoretically, it must at least be admitted that the relations of the soma and the germ plasm are different. Since the plant metamers are derived one from another by vegetative propagation, it is more obvious with plants than with animals that the differences of successive metamers are differences in the expression of the characters, rather than differences of transmission.

The metamers are morphological equivalents of each other, and may be formed in long series of closely similar individuals, or may show successions of strikingly different types of metamers, with or without intermediate forms. The size and form of a plant are determined largely by the number of vegetative metamers, which may be controlled to a great extent by the conditions of growth. Thus the characters of a particular plant are determined not only by the content of heredity transmitted through the seed, but may be modified greatly during the course of development. The conditions of existence may vary so much as to interfere with the normal sequence of expression of the characters of the different classes of internode individuals and the plant may be unable to attain a complete development of the different types of metamers.

The higher animals are much less exposed to environmental interference in development, or to observation of the succession of characters in the stages of development. Many of the lower animals are as obviously metamerous as plants, like the worms, millipeds and centipeds, and such groups may afford cases of approximation of characters in different parts of the body, as in plants. In the centiped genus Newportia, for example, the last pair of legs is very slender and many-jointed like the antennae, and the same might be said of the delicate, closely articulated extremities of all of the legs in another family of centipeds, the Scutigeridae.

Sepals Reduced in Regular Sequence

As examples of metaphanic variation the abnormal rose sepals afford material for study in detail. Even among the sepals of the same calyx, the differences are not to be taken as merely arbitrary or chance variations, but show successive stages of reduction and transformation of the characters. Probably these stages correspond to a process that occurs normally in changing the characters from those of the expanded form of the foliage leaves to the simplified and reduced form of the sepals.

To appreciate that there is a regular sequence of reduction in the metaphanic sepals, account must be taken of the morphological fact that the leaves of plants are arranged generally in spirals, and that the spiral arrangement continues among the sepals as corresponding to leaves of successive internodes of the branch. The internodes between the leaves may be so shortened or suppressed that the leaves or corresponding organs appear as a circle or "whorl" as in a calyx, corolla, or involucre, but the members of such whorls still represent structural successions, as may be seen in these metaphanic sepals.

Considering the largest sepal as the lowest of its whorl, the relation with the other sepals appears quite regular. In each case the sepal next in size is nearly opposite to the largest sepal, while the third sepal in size has a similar relation to the second sepal. The smaller sepal that projects between the first and second is the fourth sepal of the spiral series, while the fifth sepal stands between the second and third. To reach the five sepals in the order of reduction in size, two turns around the center are indicated corresponding to a 2/5 spiral of normal phyllotaxy.

Enlargement of Stipules

In comparison with the two subfloral leaves of Figure 11, the enlargement of the stipules of the metaphanic sepals is a striking feature, especially in the large calyx of Figure 10, where the stipules not only are very large but are strongly curved.

An explanation of the enlarged stipules may be found in the stipular nature of the sepals, meaning that the normal sepals correspond morphologically to the foot or stipular portion of the normal leaf with the blade of the leaf suppressed, or reduced to a minute apical appendage. Thus the enlargement of the stipules in the metaphanic sepals is to be taken as representing a stage in the normal transformation from the foliage leaf to the stipular form that is reached in the normal sepal.

Metaphanic Rose Sepals
Figure 10

Transition from leaf to calyx
Figure 11
Showing intermediate expression of characters, with leaf-characters and calyx characters combined in the same organs. The largest metaphanic sepals are larger than the first leaves below the calyx, as shown in Figure 11. Intermediate stages between stipules and pinnae of compound leaves are shown in the metaphanic sepals. The numbers show the phyllotaxy of the sepals. The two leaves are from the same stem, immediately below a metaphanic calyx, showing that the usual reduction in the size of the leaves below the flower had not taken place, or only to a slight extent, under the conditions that induced the metaphanic variations. Examples of gradual reduction of leaves may be seen in Figure 13.

Reduplication of Stipules

Another feature of the metaphanic rose sepals is that some of them show distinctly two or three pairs of stipules, or of stipule-like organs, in some cases so closely alike that their morphological equivalence would hardly be questioned. The intermediate stages indicate that the lower pair of these structures corresponds to those that function usually as stipules, while the upper pairs correspond to the pinnae of the compound leaves.

Thus an indication is afforded that the compound leaves of the rose family have developed by reduplication of the foot or stipular element, as already indicated for the leaves of the walnut family, in the JOURNAL OF HEREDITY for May, 1923. The rose family retains the first pair as stipules at the base of the leaves, while in the walnut family the stipules have become obsolete in the fo1iage leaves, though still appearing in the bud-scales. In these groups it appears that the lateral pinnae of compound leaves correspond morphologically to the stipules of simple leaves, instead of the compound leaf type having been developed through subdivision of simple leaves. But compound leaves in other families may have followed other courses of development.

Middle Sepals Unsymmetric

It is worthy of note that the third sepal of each calyx is usually unsymmetrical. As the photographs show, the stipules are distinct on one side of the third sepal, but on the other side are completely suppressed. This seems to be true in all cases where the stipular elements are developed as separate lateral appendages. Thus with respect to the stipules, each series of five sepals divides exactly in the middle. Two of the sepals have the stipules distinct, two have the stipules completely fused, and the middle sepal has the stipular element separate on one side and fused on the other. In all cases the appendages are found on the side that is lower down on the spiral of the phyllotaxy, in the direction of the sepals with the free stipules.

With the sepals differing so much in size, as shown by the photographs, it seems the more remarkable that the change from separate to suppressed stipules should appear so definite, and that it should take place at the middle point of the series of five sepals in each calyx. That this specialization is so definitely established may be connected with the overlapping of the sepals in the bud, so that the third sepal is exposed on the lower side and covered on the upper side, although the presence or absence of the separate stipules undoubtedly is determined at a very early stage of development, far in advance of the difference of exposure.

Although the reduction in size of the sepals is gradual, the suppression of the stipules, or fusion with the foot element, still appears as an abrupt or definitely contrasted character. The facts might be interpreted by supposing that the lower sepals have been replaced by leaves of gradually reduced sizes, but on this basis the middIe member of each calyx would need to be reckoned as a composite structure or chimera—leaf on one side and sepal on the other. Such chimeras do occur, and may be taken as further examples of abrupt changes of characters in the course of development, changes so abrupt that the transition occurs during the formation of a metamer, rather than between metamers.

One of the sepals of the calyx at the bottom of Figure 12 shows a chimaeroid anticipation of petal characters, in the lobe-like expansion of the base on one side, which had the texture and color of a petal. It is notable that this variation occurred in the fourth sepal of the series, and that the fifth sepal was normal. Thus there was only a transitory deviation, as though the petal characters were switched temporarily into expression when the rudiment of this particular sepal was formed.

Variations like Chimeras

The distinction between a metaphanic chimera and other metaphanic variations would lie chiefly in the abruptness of the change of characters. A transition may be so gradual as to cover the formation of several intermediate metamers, or may be so abrupt as to occur at a definite point in the formation of a single organ. While the internodes are individuals, they are not completely autonomous, since they are subject to changes of characters, even after their development has begun. The change from leaf to sepal, as shown in Figure 11 still is an abrupt change, notwithstanding the unusual enlargement of the lower sepals, while the changes shown in Figure 13 are much more gradual.

Examples of more abrupt metaphanic variations of stamens, petals and sepals were found together on a tree of Texas Prolific almond, at Sacaton, Arizona, March 6, 1921. Some of the stamens were expanded to form supernumerary petals, some of the petals showed sepal characters, and some of the sepals showed petal characters, either in median segments of the organs or on one side. Several of the metaphanic petals had midribs of thickened deep purple or green tissue pubescent on the back like the sepals, and these modified petals persisted like the sepals after the normal petals had fallen. Some of the metaphanic sepals had a large wedge-shaped petaloid expansion in the middle, while the lateral segments of the same organs showed normal sepal characters.

One of the modified stamens had a rather narrow, contorted, petaloid white blade on one side, while the other side was represented by a small, yellow abortive anther cell attached at the same elevation as the anthers of the normal stamens, not at the end of the blade, but at the end of a shorter thickened margin that doubtless represented one side of the filament of a normal anther. This abnormal petal-stamen is a chimera, half petal and half stamen but both halves with modified intermediate expression of the characters. Such an abnormality suggests that the expanded portion of the petal corresponds to the anther of the stamen while the narrow base or petal-claw corresponds to the filament. Also the occurence of metaphanic variations of the three organs together is interesting as a suggestion of homology and possible derivation of all of the expanded organs, petals, sepals and leaves, by a process of sterilization of stamens.

Reduplication of Stipules
Figure 12

Gradual Changes of Characters
Figure 13
Many of the metaphanic sepals have three pairs of stipule-like organs, the lowest pair corresponding to the normal stipules and other pairs representing the lateral pinnae of the normal foliage leaves. Observe that in each calyx one of the sepals, which is the middle member of the series, has stipules distinct on one side, while the other side is without stipules. Reduction of leaves to bracts showing gradual stages of enlargement of the foot or stipular element of the leaf and of suppression of the blade element, to form the small sepal-like bracts that are found below the calyx in some varieties of roses. The gradual reduction of the leaves, even in advance of the formation of the stipules, is in contrast with the abrupt change of characters shown in Figure 11.

Expression Relations of Characters

From the facts of metaphanic variation it is plain that the characters may be studied in other relations than those of transmission, seeing that plants with the same content of transmission may still develop quite differently. Differences may be shown not only in size or in other direct effects of the environment, but also in the sequence of development of the various kinds of metamers and in the transitions from one type of metamer to another. To know how these expression-relations are modified by the conditions of existence may be just as important as to know how characters are transmitted.

How the characters are represented in the germinal material is still obscure. To determine that the characters are represented in the chromosomes is a remarkable advance of scientific knowledge, but does not give us functional ideas of the nature of the characters or of their relations to each other in the sequence of development, or in the course of evolution.

Metaphanic Almond Flowers
Figure 14

Enlarged Views of Almond Flowers
Figure 15
Showing modified petal-like stamens, sepal-like petals and petal-like sepals. First flower in upper row a corolla with 6 petals, one of these a modified petal-stamen overlapping a sepal-petal with a deep apical notch and a dark midrib. Second flower with two normal petals attached and small petal-stamen as described in text. Third flower with normal petal and deeply notched sepal-petal with green midrib. Fourth flower a fasciation or twin with 8 sepals, two sepal-petals and one nearly normal petal but remaining attached while the others had fallen. Flowers and separate organs in lower rows show numerous examples of sepal-petals and petal-sepals, some of the latter as half-sepals, petaloid on one side, some as segmental chimeras with wedge-shaped expanded sections of white petaloid tissue in the middle, between sections of normal sepal-tissue, green or deep purple in color and pubescent on the surface. Enlarged views of some of the flowers shown in Figure 14, showing in greater detail some of the metaphanic variations. Normal petals and sepals are shown, and forms suggesting sepal-petal chimaeras. Such "inter-organ hybrids" between petals and sepals are good examples of metaphany, a type of variation possibly of great evolutionary significance. The difference in these parts must be assumed to be due to differences in expression, rather than to differences in heredity, and the ability to form such intermediate variations may afford the organism a certain latitude of experiment in making adjustments to the environment.

Since many characters may be transmitted without being brought into expression, the expression-relations must be considered even for an understanding of transmission. Differences of expression-relations are recognized in cases of dominant and recessive characters, but these have a wider significance in the field of development and in the study of evolution, Apparently it is necessary to work from the side of expression to gain more concrete ideas of the nature of the characters, and of their relations to each others. The metaphanic variations are not to be explained merely by saying that they are due to misplacement or "translocation" of the characters. The idea of translocation is misleading, as tending to associate the phenomena with the transmission of characters, instead of recognizing that the expression is changed, rather than transmission.

It is significant that the characters are changed in the course of plant development, so that the process of development is a more or less definite sequence of changes of characters. The changes of characters that appear in successive metamers are the phenomenon that led Goethe to recognize the "metamorphosis" of plants. Goethe saw clearly that the internode individuals were structural equivalents, and became convinced that the differences which had developed between the different kinds of internodes were essentially of the same nature as the differences between species.

These analogies are definitely stated in Goethe's prose writings as well as in his poem on plant evolution, but such morphological conceptions were not understood by Goethe's literary commentators, and even among present-day biologists are strangely overlooked. No better evidence or illustration of evolution has been presented by later writers on the subject.

The Problem of New Characters

One of the difficult problems of heredity and evolution is to understand how new characters originate. In this direction little progress seems to have been made. The more we accustomed we are to think of heredity as a definite mechanism for repeating the same characters in each generation, the more difficult it is to entertain the idea of suddenly implanting a new character, to say nothing of adding several new characters at once, as assumed by those who believe in abrupt transformation into new species.

The idea of defect or derangement of characters, or of characters failing to come into expression, is well established among students of genetics, and undoubtedly applies to many variations of the mutative, Mendelizing type, but it does not account for new characters

Variations occur in all groups of organisms, and have been studied extensively, but with no agreement among specialists as to which classes of variations contribute to evolution. Although the general idea of evolution is accepted by most biologists, the lack of a clear understanding of the nature of the evolutionary process leaves the question still in debate. It is easier to elaborate arguments about evolution than to become familiar with the facts and difficulties of the problem.

Darwin found reasons for believing in very slow and gradual changes of characters, and ascribed such changes to the influence of natural selection in accumulating slight, imperceptible differences through long periods of time. But this Darwinian theory, that selection is the cause of evolution, though still credited by many, is held less firmly and may be abandoned entirely.

Experiments have not shown that selection has any power to induce variation or to bring about the formation of new characters. Evolution is a creative process while selection is a process of elimination. Selection functions as a "safe-guarding principle," by weeding out the defective and inferior that otherwise might hamper or restrict the evolution of the species, but selection does nothing to the survivors, to cause them to change, or to induce changes in their offspring.

*Evolutionary Inferences from the Diplopode, Proc. Entomological Society of Washington, 5:84. 1902;
   Aspects of Kinetic Evolution, Proc. Washington Academy of Sciences, Vol. VIII. pp. 197-403, 1907;
   New Tropical Millipeds of the Order Merocheta, with an example of Kinetic Evolution, Proc. U. S. National Museum, 40:451, 1911.
It is easy to see that selection may prevent or deflect evolution, if some variations are removed while others are retained, but this affords no basis for the assumption that selection is an actuating cause of changes that go on. The whole argument for and against materialism, elaborated on both sides from the assumption of an environmental cause of evolution, is irrelevant. Evolution does not depend upon selection or other environmental causes. The evidence that bears upon this question tends to show that evolutionary changes of characters are spontaneous within the species, so that each species can evolve independently, in the direction of adaptive or non-adaptive variations.*

Abrupt and striking differences appear among the members of species, and many biologists are still engaged with the idea that such saltatory variations or "mutations," as they are now called, are the means of changing one species into another. Darwin studied many of the abrupt variations, or "sports" but considered them as abnormalities rather than as constructive steps in evolution. It remained for Professor de Vries to rehabilitate the idea of sudden changes of characters as giving rise to new species. Interesting facts were presented and many readers were convinced, but no general acceptance of the transformation theory is not expected.

*Evolution not the Origin of Species, Popular Science Monthly, March, 1904;
   Factors of Species-Formation, Science, March 30, 1906;
   Evolution without Isolation, American Naturalist, 42:727, November, 1908.
The emphasis placed by Darwin and de Vries on the origination of new species is misleading, since the differentiation of species is not necessarily connected with the formation of new characters. Species may be separated without producing new characters, or new characters may be established in a species without any subdivision of the group taking place. Changes of the characters of a species may be more significant from the standpoint of evolutionary progress than a separation into a larger number of species.*

It seems remarkable that so many attempts have been made to explain evolution without a preliminary effort to gain clear conceptions of the nature of evolution, as a process of change in the characters of species. The species as they exist in nature, not only must be considered as products of evolution, but also as showing the condition in which evolution takes place. In other words, the nature of species affords the best indication of the nature of evolution.

For the purposes of evolution and heredity it should be recognized that the natural species is a group of diverse individuals united by interbreeding into a continuous network of descent. It is in such groups of organisms that evolutionary progress must be supposed to take place, rather than in uniform, "pure-line" groups which many writers have assumed to be species.

The study of metaphanic variations may help us to understand how some of the mutative diversities among the members of species may contribute to evolution, though most of the abrupt changes of characters are abnormal, as Darwin recognized. Indications are afforded by the metaphanic characters that wide changes of external features may result from extremely slight changes in the internal mechanism of heredity. It need not be supposed that the transmission of the characters is changed, but only that the expression of the characters is modified, so that the extent or sequence of development is affected.

The same inference regarding variations as changes of expression could be drawn from the many cases where widely different species and even genera are still able to hybridize, and from the many cases where the adult characters of some species appear as juvenile characters of related species. Thus it appears that many of the differences between species are of the same nature as the diversities that are found among members of the same species, and also are like the variations that occur in equivalent parts of the same organism. Three kinds of variations, specific, heteristic and metaphanic, may be considered as changes of the expression-relations of the characters, without supposing that the transmission of the characters has been changed.

Instead of such differences of expression being too small to be perceived readily, the difficulty is rather that they are so numerous and versatile as to seem intangible, so that it has been possible for Darwin and many other writers to confuse the normal genetic diversity, or heterism, of species with fluctuations in size and other responses to environmental conditions.

That some of the metaphanic vanations may be called forth by special environmental conditions does not render them of less interest in heredity and evolution, but rather of more interest. Liability to deviations in development is a quality of the hereditary constitution, in contrast with greater regularity of expression of other characters, or in other lines of descent. Though many of the more striking metaphanic variations are obviously abnormal, some may not be defective, or may be of value as adaptations, so that they would be favored by selection and become established in the species. How many of the heredity variations can be interpreted as metaphanic, or due to modified expressions of the characters, remains to be determined.

Meehan: Stipules as Petals (1887)

Rose 'Marquise Boccella' with shoot growing from base of leafy sepal (left), and leaves in place of sepals (right).

Petaloid anthers in Camellia (left) and Azalea (right).

Hemerocallis Kwanzo (left) and a rose (right)
with metaphanic petal-anthers.
Metaphanic variations are easy to find. Double flowers frequently show petal-stamen metaphanic hybrids. While photographing some examples on a rose, I noticed a few petaloid stamens in a flower on an otherwise single-flowered azalea.

Agapanthus sometimes have spathes colored blue like the flowers. These may be considered metaphanic chimeras, but are reminiscent of other plants with colored spathes: haemanthus, callas, anthuriums and other aroids.

A particularly interesting and informative metaphanic anomaly was found in the North Caucasus, and described by the great Russian plant breeder Ivan Michurin—the Vegetativnaya pear. Michurin wrote:

     It was ascertained that in the spring, like all other varieties, this pear blooms and bears fruit of medium size and of fairly good flavour. After the fruit ripens, in the first half of July, the tree blooms a second time and simultaneously, the buds on the young shoot give rise to small stalklike formations which, in their turn, bear several buds at the tips crowded close together; from these buds leaves begin to develop.
     As these leaves grow, their petioles gradually and simultaneously begin to swell, and by the end of August and beginning of September, when the pears from the second blooming are already ripening, the fruit that was formed vegetatively from the petioles also acquires pear shape; and by this time the petioles are quite imperceptible, because it was from their simultaneous thickening on the fruit stalks that these fine pears, which for size and flavour cannot be distinguished from fruit that had set at blossoming, were formed.
     The leaf blades of the petioles from which the "asexual" pears are formed do not disappear like the stalks; they remain in the place of the calyx, gracefully protruding from the top of the vegetative fruit in the form of a large bunch of rosettes (see Fig. [131]).

This odd pear, understood as a metaphanic variation, reveals that the fruit of the pear corresponds approximately to the petioles of leaves. Cook already found that the filament of stamens correspond to the petal-claws, so we may suppose that pears develop from the corresponding portion of the sepals. Thus, in the Vegetativnaya pear the usual floral parts (sepals and petals) had become metaphanic hybrids resembling leaves at first, then altering.

Evolution of Plant Structures

Cook Bibliography