Resumo: |
1) The present paper deals with variations in the structure of the inflorescences of the Maydeae Americanae, especially of the ear of Zea Mays, which came under observation during our genetical studies during the last 10 years. Care was tauken to include only such structures which were observed frequently and had in general an aspect which could not be described as a pathological abnormality, but rather as a reversion to a phylogenetically primitive type or as a new form. The aberrations may be summarized as follows: a) Individual flowers. Eeversion to the hermaphrodite state of the usually unisexual flowers in Maydeae were described. Such flowers are however already so well known that no details need to be given. b) Many flowered spikelets. The spikelet in all the American Maydeae are normally two flowered. In forms of tunicata, obtained by selection, spikelets were obtained with up to three perfect flowers, all producing well developed kernels (Fig. 4-6), with a maximum of five flowers. The structure of three-flowered spikelets is explained in the diagrams Fig. 2 and 3. Owing to the organization of the flowers, the embryos of the first and third are turned towards the base of the inflorescence and the embryo of the second flower towards its tip. The embryos in two-flowered epike-lets are normally turned to the base (first flower) and to the tip (second flower). (Cutler, 10). Cases were observed of three flowered spikelets with one abortive flower, both the remaining embryos, produced by the first and third flower, turned towards the base of the inflorescence. c) Number of spikelets per alveolus. In maize and in the male parts of the inflorescences of Euchlaena and Tripsacum two spikelets are formed per alveolus, while in the female parts of the latter genera one spikelet is present, though occasionally exceptions are found. Collins (8) explained the appearance of the double spikelet as remainder of a small spikelet branch which has been lost during evolution. The normal structure and that of a spikelet branch is shown schematically in Fig. 3, e-f. In the tassels of several tunicata plants small and regular branches of spikelets appeared in the alveoli, their form depending upon three factors: a) number of spikelets, b) length of the internodes of the spikelet branch, c) length of the individual pedicel of each epikelet, as may be seen in the fotos of Fig. 4-5. An increase in the number of spikelets¹ per alveolus in the ear has been observed frequently in material of two sources: deeendants of the hybrid maize x teosinte and Paulista Pointed Pop Corn. The descendants of the cross Zea x Euchlaena have very frequently only two rows of alveoli, well separated from each other. This separation is accentuated still further by the development of a scaly outgrowth from the rachis, as in Euchlaena, which limits each alveolus at its base and on both sides. The segregates may have these scales not only well developed, but they are also frequently colored, as is the eases representad, with black dots. Thus the identification of each alveolus is very easy and fig. 7 shows an ear with two rows of well defined alveoli, each with three spikelets or three mature kernels each. The ears of Paulista Pointed Pop Corn offer excellent material since here the rows of alveoli are very clearly defined and separated by deep ridges, the double rows of kernels being very salient. The kernels correspond always, in the ears used, to single spikelets as shown by their glumes, well developed in this variety, and the position of their embryos. In Fig. 9, a, the deep ridge separating the rows of alveoli on the left and right is clearly seen as a series of black deep shadowns in the centre of the photo . In Fig. 9,b we have the same ear photographed from another angle. One row of alveoli is seen in the centre, in clear contrast against two strips of black paper inserted in the two lateral ridges. This row of alveoli contains very clearly two rows of kernels or spikelets in its upper part and many irregular rows in the lower half showing the increase of spikelet number. Fig. 9 c-f represents another ear with five rows of alveoli, which are very distinct in the upper part of the ear. The ear was broken into three parts and the cross, sections' phographed. The upper two thirds with a regular row arrangement (Fig. 9 c-d) show clearly the five rows of alveoli and the 10 rows of kernels. The lower part (Fig. 9, e) shows still the pentagonal form of the raehis and the five rows of alveoli may still be distinguished. But the increased number of spikelets, which vary from 2 to 4, do not permit a regular arrangement of rows of kernels within the regular rows of alveoli. The appearance of small branches in the alveoli of mutants of maize is not anything really new, but as far as I know an increase of the number of sessile spikelets in the ear from two to three which was observed in several descendants of the Zea-Euchlaena cross and its increase to three or four in the basal region of ordinary corn ears has not been registered before. The explication for this form of increase is easily to be drawn from an inspection of the photos in Figs. 4 and 5. where we find various stages in the reduction of branches with many spikelets to two, three or four sessile spikelets. d) Number of rows of alveoli. There can be little doubt that the basic number of rows in the- Maydeae Americanc is always two as shown by all inflorescences is Tripsacum and Euchlaena and by the tassel branches in maize and the branches in Zea "ramosa". The increase of rows in the ear and the central spike of maize has been the object of a very extensive literature. Mangelsdorf and Beeves, (15), gave a very detailed review), and I explained this increase as a simple interpolation of new rows, such as occurs in succulent plants and can be easily seen for instance in different species of Cereus. The correctness of this interpretation can be shown directly in ears of maize which have two rowd at the tip and more numerous rows at the botton and which appeared in my material of Brazilian Early and Brazilian Sweet Corn, extracted from multiple crosses between Brazilian, European and North American corn, and with much more frequency among the descendants of the maize-teosinte cross. Three such cases are reproduced in Fig. 10. . The situation is represented schematically in Fig. 11. The diagram Fig. 11a represents the primitive from with two rows of alveoli, each with two rows of kernels. The interpolation of two new rows (Fig. 11b) causes a disturbance of the equilibrium which forces the two original rows to become dislocated, as shown in Fig. lid, thus resulting finally the normal ear with four rows of alveoli or eight rows of kernels, all equally spaced. The interpolation of a single row of alveoli causes a similar disturbance (Fig. 11c) which causes distortions untill all rows are equidistant (Fig. 11e). For some reasons, still unknown, the ears remain twisted in the latter case with three rows, while in the former a twisting occurs only in the transition zone. These drawings however are not abstractions, but represent the situation actually observed and represented by the Fig. 10, a-b e Fig. 10, c-d, with in increase from two to four rows and by the Fig. 10, e-f with an increase from two to three rows. The two first ears, photographed each from two different angles, show clearly that the two rows of the top region become neighbours at the botton, while the new rows are interpolated on the opposite side of the rachis. The last ear shows the permanent twist ra the lower third of the ear, and it may be added that all ears, wich I have observed so far, with three rows of alveoli are always twisted over their whole length. I postulated furthermore that originally the alveoli followed an alternating spiral arrangement or a phyllotaxy of 1/2, while the interpolation of new rows abolished this spiral and substituded a. longitudinal row arrangement, again as in Cactaceae and other succulent forms. The fotos in Fig. 10 show that this is actually the case. The arrangement of the alveoli in the two-rowed part of the ear follows the spiral (½), while the spiral is completely lost in the many rowed region. The rows may be twisted in ears with an uneven row number, but this twist has nothing to do with the original spiral of the alveoli, as may be seen clearly in Fig. 10, e and f. It may be mentioned that the situation in the latter ear is still more complicated by the fact that the alveoli contain frequently three, instead of two spikelets. e) Asymmetry of the inflorescence. In the male inflorescences of all the Maydeae Americanae a pronounced asymmetry of the rows may be observed which are not localised on the sides of the rachis but are pushed towards its ventral surface. In the female parts of Tripsacum and Euchlaena on the other side, the spikelets are alway on oposite sides, while nothing can be said with regards to the many rowed corn ear and the central spike of the tassel. I have no doubt that the original type of maize with only two rows of alveoli was asymmetrical, which seems to be very clear when inspectionig two rowed ears or too rowed parts of ears of mutants. It is very interesting to note that there is a segregation in the descendants of the Zea-EucMaena cross with two rowed ears, which may be symmetrical (Fig. 8, c-d) or asymmetrical (Fig. 7, Fig. 8, a-b). The symmetry is independent from the number of rows of kernels which may be either one or two per alveolus. Besides this asymmetrical arrangement of the rows of alveolus, there may appaer in maize on asymmetry between the rows of spikelets: a regular combination of one row of sessile female and pedicelled male spikelets. Such a situation in the tassel branches of tunicata and tassel seed is shown in Fig. 13, a and b, and that of an ear with 4 rows of alveoli in Fig. 13, c e d. f) A detailed and comparative study of the ramification of inflorescences in the three genera of Maydeae Americanae showed an extreme degree of uniformity, which apparently has not been noticed to this extent before. The terminal inflorescence, of main shoot or branches, has essentially the same structure: there is always a central or main spike which may be branched at its base. In Tripsacum, these branches follow an alternate spiral arrangement while in the tassel of maize and teosinte and in branched maize ears they form one or more whorls. The alveoli, even in many-rowed maize, are always arranged in two rows on all these branches. Below the inflorescence and again in all three genera there are several sterile leafs which may be more or less transformed into husks and which only very excepcionally produce branches in their axils. The branches, produced by the lower leafs, terminate again in an inflorescence which may be branched or not, followed first by sterile leafs or husks and than by fertile leafs bearing again branches of higher order. In Tripsacum australe all branches of any order are essentially alike, except for a progressive reduction of the branching of the inflorescence. Branches of third or fourth order bear simple terminal spikes. In Zea and Euchlaena the lateral branches of main stalks form a compound which we may call with Goebel (11) an anthocladiun. Its structure is explained by the two fotos in Fig. 14 which were token from two descendants of the teross Zea x Euchlaena. These authocladiums have a terminal inflorescence which may be a simple ear (Fig. 15b), a slighty (Fig. 14) or an extremely branched inflorescence, carrying male and female spikelets on its four rowed branches (Fig. 16) or which may finally be a normal tassel (Fig. 15a). There follow a few sterile leafs, forming the individual husks of the terminal inflorescence and than fertile leafs or husks which are protecting the rest of the anthoeladia. In their axils there are again branches (n+2) with terminal inflorescences, always represented by an ear, protected by their individual and sterile husks, followed by other husks which protect both the branch (n + 2) and its higher derivatives. The authocladium of Fig. 15 carries branches up to those of order (n + 4). The difference of Zea and Euchlaena is only of quantitative nature. In Zea generally there is developed only the terminal ear of the lateral branch of order (n-f-1) and only rarely some ears, terminal of the order (n+2) which may be called also lateral to the branch (n+1), as shown by Brieger (6). All leafs are transformed generally into husks. In Euchlaena, on the other side, the terminal inflorescence of order (n+1) is generally a tassel and this branch carries almost normal leafs. Those of the lower nodes produce in their axils the antoeladia, with the branches (n+2), (n+3), etc. which together form the cluster of ears. Every ear as a terminal inflorescence of its branch is protected by individual husks, besides the basal husks, as described above, which protect each branch and all its laterals of hingher order. In the descendants of the Zea-Euchlaena corn an interesting detail was found, and the following three types were established: a) maize type derivatives where all inflorescences had a high number of rows of alveoli, b) teosinte-type derivative® where all inflorescences had only two rows of alveoli, and c) intermediate derivatives where only the main terminal inflorescence of order (n + 1) was many rowed and some-timep also inflorescences of order (n + 2), especially the prophyllary branch, all other inflorescences beeing two-rowed. These complicated relations obscure frequently the rule established by Langham (14) about the correlation between row number in the lateral ear and the central spike of the tassel. These observations lead to a revision of the old hypothesis of Ascherson (2) about the origin of the many-rowed ear by a fusion of the ear clusters in Euchlaena. Weatherwax (20) has given already one very decisive argument against this hypothesis, explaining that a fusion must result in the forming of an ear where the number of kernel rows should be a multiple of four and that of alveoli a multiple of 2. Another difficulty is that the original hypothesis would require a complete disappearance of all individual husks, and no signe of this has ever been observed. Langham (14) who tried to revive the hypothesis was probably mistaken, interpreting as a fusion, structures which were only branched terrninal ears, which its branches forced almost into the main spike by the pressure of the involving glumes. 2) After describing these deviations form the normal type of maize ears and inflorescences, its genetical basis was discussed. First it was thought that definite genes were involved, such a the ramosa-gen, the dwarf-factors which frequently cause the appearance of perfect bisexual flowers, the tassel seed factors, etc, and finally the tunicat factor which has the most¹ extreme range of phaenotypie effects as shown in several previous publications (Brieger, 4, 5, 6, 7). Buth thei studies of tunicata maize had already shown that the modifier gens are at least as important as the main determing gen. By selection I succeded to increase to a extreme degree their effects on the tassel. Crossing these derivations of Paulista Pod Corn to Euchlaena a complete change of the situation was brought about. The effects on the tassel disappeared immediately in the Fl with Guatemala teosinte while in those with Mexico teosinte the glumes in the tassel remained shighthy enlarged. On the other side, changes oecured in the ear structure unknown to pure maize, as illustratad by Brieger (6). The descendants of these crosses showed that the specific determiner genes of maize are dispensable and that a recombination of Zea and Euchlaene modifiers a is completely sufficient. I described in the present paper only variations in the structure of inflorescence which appeared both in descendants of the hybrids and in pure maize. No similar vairiations have been observed in Euchaena or in Tripsacum, the third species of the group. We may thus conclude that these variations were all inherent to corn, hypostatic recessive to the many gens accumulated during domestication and reappearing again after the substitution of these genes of domestication by the wild type modifiers derived from Euchaena. 3) The final conclusions of the preceding chapter leads us to a reconsideration of our concepts about the structure of the wild ancector. I discussed the problem in several previous publications (4, 6) and shall repeat the main conclusions without entering into a detailed discussion. Cultivated maize is derived from a wild grass which had simple ears with two rows of alveoli, grains protected by large, pointed and horny glumes, due to the presence of the gene Tu, and an abscission layer at the base of each spikelet. Thus the grains fell off when mature with their protective glumes, separated from the rachis which was thin, flexible and not brittle. This grass happened to cross with some form of Tripsacum whith its grains mainly protected by special scales, having relatively small glumes due to the presence of the gen tu The rachis of this species is always horny, brittle and breaks on maturity into parts at each node. In natural descendants of these hybrids segregates appeared which weTe homozygous tutu and thus had small glumes, while maize factors suppresed the development of the scales. Both abscissions layers were mutually supressed, that at the base of the spikelet from maize and the other, from Tripsacum at each node of the rachis. We now may add more informations about the probable structure of the original inflorescences. The lateral inflorescences were asymmetrical and contained in each of the two rows of alveoli one row of sessile female or hemaphroditic spikelets and another row of pedicelled male spikelets. The very pronounced protogyny of flowers in the same inflorescence permitted that the inflorescences remained within the husks or within the isheaths of only slightly altered leaves during the female phase, emitting only the silks. During the following male phase, the inflorescences! were pushed out of the glumes by the lengthening of the internodes, thus permitting the shedding of the pollen, white at the same time the glumes of the already fertilized female flowers had becone already big and horny, protecting the kernels in the milkystage. The terminal inflorescences may have been branched with one or more whorls of branches at its base, and the number of branches was less in the lateral inflorescences of higher order. There may have been already some more pronounced differentiation of terminal and lateral inflorescences with a predominance in the former of male spikelets. There exists a possibility that the spikelets were many flowered and thus this wild grass did not really belong to the Andropogoneae, as already supposed by Harshberger. It seems now possible to make a drawing of this hypothetical wild type, but I think it even possible to find eventually among the descendants of the maize-teosinte cross most of these characters in combination, thus permitting to take a photograph. 4) The results described opened the way for drawing still more general conclusions about the mechanism of evolution, by combining two sets of facts with the new results obtained. In the first place, we may mention the conclusions at which Agnes Arber (1) arrived in her detailed study on the evolution of the order Gramineae. She stated that there exists a) "inherent phylogenetical trend" for reduction and sterility in inflorescence and flowers, and b) a large measure of repetition and of parallel evolution in different systematic groups. In the second place I may refer to the general hypothesis about the role of a modifier shift in evolution which I have developed (Brieger, 6). My observations started in 1929 when I observed in the cross Nicotiana Sanderae x N. Longiflora that a gene, without effect in the latter species, acquired a semilethal effect in the Fl hybrid. An identical case was reported a year later by Hollingshead. In subsequent studies, I have been able to show experimentally the importance which a shift or rearrangement of the modifier complexes may have, and which may not only affset the dominance of certain characters, a fact áÍLTeady very well konw by many authors, but may cause the appearence of completely new phaenotypes. Such a change in the modifier complex may be brought about by selection, artificial or natural, following a gen mutation which opens new phaenotipic possibilities or simply after a change in the direction of selection in material sufficiently heterozygous. In the present paper I have shown that there exists in cultivated maize an enormous amount of "inherent possibilities" which may become reality after changing the composition of the modifier complexes. Thus a "parallel evolution" may become started, or even a "reversion of evolution". It may be sufficient to cite the appearance of plants with two rows of three spikelets each, such as in Hordeum, or of many flowered spikelets as in the Paniceae. Thus the facts reported here offer more support for the hypothesis on the importance of the modifier shift in evolution which seems to me a very efficient corollary to the mutation theory and helps to overcome at least one serious difficulty of the latter. Itis extremely difficult to believe that the accumulation and selection of disorderly chance mutations may cause an orderly evolution, Besides, the mutation theory is incompatible with such concepts as that of "inherent trends". But since the latter undoubtedly exist and since evolution is not disorderly, as shown by a very large number of comparative studies of larger taxonomical groups, an explanation for this orderliness must be found. I believ that the hypothesis of the modifier shift offers the necessary explanation, which is well in accordance whith experimental evidence, and with our concepts of physiological genetcs.
|