In this unique and comprehensive article, Dr. Cusset presents a historical review of ideas and theories on leaf development, a detailed description and critique of the classical model, a discussion of Jeune's model, and finally his own synthetic model.According to Cusset, the classical model, which has been widely adopted in textbooks, reference books, and research, postulates that (i) leaf morphogenesis results from the successive activity of apical, marginal (or submarginal), and intercalary meristems; (ii) cell files arise from submarginal initials; (iii) the cell is the fundamental unit of morphogenesis. Cusset documents the shortcomings of these assumptions by discussing evidence from publications that deal with modem techniques ranging from the quantitative analysis of mitotic activity and chimeras to marker experiments and clonal analysis. All of these different approaches have led to the same conclusion: for nearly all taxa studied so far, there is no evidence, either in terms of mitotic activity (e.g., mitotic index) or in a functional sense, of apical or marginal meristems. This means that apical or marginal meristems cannot account for the elongation and lateral extension of the leaf blade. Rather, the available data support the following conclusions.(i) The young leaf primordium is fully meristematic, i.e., there is generalized mitotic activity throughout the primordium. Gradually the mitotic activity is restricted to the base and thus the leaf develops through the activity of an intercalary meristem. Both the orientation and the frequency of cell divisions indicate that the leaf margin plays only a minor role in the growth of the lamina. In only a very few cases (e.g., the petals ofNicotiana tabacum), and then in only relatively late developmental stages, has a maximal mitotic index been observed directly at the margin. More commonly, a peak of mitotic activity has been noted in a submarginal region. However, this region is at a considerable distance from the margin, and the increased mitotic activity, which is often not statistically significant, occurs in only some developmental stages. In summary, one may conclude that instead of distinct meristematic entities, the developing leaf shows a patterned continuum of meristematic activity.(ii) The so-called submarginal initials contribute relatively little to the lateral extension of the leaf blade. Hence, it is questionable whether they should be called "initials."(iii) In general, cells do not appear to be the fundamental units of morphogenesis. As pointed out long ago by De Bary: "the plant forms cells, not cells the plant," thus underscoring the need for a more holistic approach.In addition to the preceding conclusions, Jeune's model (developed in Cusset's laboratory) incorporates the notion of growth centres ("centres générateurs"). These are defined as fields with certain physiological properties that are transmitted to other areas of the developing leaf. Each leaf primordium has two growth centres. In compound or lobed leaves, they are responsible for the production of the lateral elements according to the following rules, (i) Each primordium of a leaflet or lobe arises at a fixed distance from the preceding one. (ii) The rhythm of their formation is constant with regard to the plastochron on the axis where the leaf is bom. (iii) As a consequence, the correlation between the number of lateral elements and the logarithm of the length of the blade primordium is linear (which confirms the exponential growth of the primordium). (iv) After the inception of a lateral element, the growth centre that gave rise to it is reconstituted. If, however, the growth centre is removed microsurgically on one side of the leaf primordium, no lateral elements are formed there. The concept of centre générateur closely corresponds to the idea of a growth centresensuWardlaw. Although Jeune's model has been confirmed for a great variety of leaves, both simple and compound, it does not fully apply in all cases, as for example in the leaves ofCastanea sativaandFraxinus excelsior. To cope with these exceptions, as well as other situations, Cusset proposes a synthetic model in which leaf blade development is modulated by the following eight internal effectors: (i) biochemical oscillations, according to the "brusselator," a model directly derived from ideas of Turing and Prigogine; (ii) the calmodulin–Ca2+balance; (iii) a distal phyllopodial organizer, based on microsurgical evidence, but rather hypothetical; (iv) a system orienting the major veins according to the model of Ferré and Le Guyader; (v) interactions between the flux of inductive substances explaining the minor venation according to Mitchison's model; (vi) a microtubule orienting mechanism which might be an internal electromagnetic phenomenon; (vii) an organizing mechanism that eventually explains the particularities of the minor venation of each species; (viii) auxin acting in a complex fashion on several of the preceding effectors. External factors, such as temperature interact with the above internal mechanisms.Although some aspects of the above synthetic model remain rather hypothetical and speculative at the present time, Cusset's proposal may serve as a stimulating working hypothesis. Furthermore, it emphasizes that the study of leaf morphogenesis not only is a technical and morphological problem of limited interest, but also involves aspects of molecular biology and fundamental theoretical and philosopical issues. From this point of view, leaf morphogenesis may serve as a model case for the discussion and elucidation of contemporary morphogenetic and biological problems.