Floral symmetry.

Monstrous flowers are curiously attractive. For years, gardeners have sought and maintained plants with abnormal numbers, types or arrangements of floral organs. In some varieties, flowers have even lost their sexual function as their reproductive parts have been modified or replaced. The scientific implications of floral monstrosities have not gone unnoticed. In 1744, Linnaeus devoted a dissertation to an aberrant plant of toadflax, Linaria vulgaris, that had flowers with radial symmetry rather than the normal bilateral symmetry. The aberrant form, which he called Peloria (greek for monster), caused Linnaeus to question whether species were as fixed and immutable as was generally assumed at that time (Linnaeus, 1744; Gustaffson, 1979). Abnormal flowers also played a major part in Goethe's theory that the different types of plant organs, such as leaves, petals and stamens, were simply variations on a common underlying theme (Goethe, 1790). He cited abnormal flowers, in which one type of organ could apparently be converted or replaced by another type, as confirming the fundamental equivalence between organs. Interest in floral aberrations gradually declined, however, towards the end of the 19th and during most of this century, as plant biologists concentrated on physiological, cellular and molecular processes. Abnormal flowers began to be thought of as rather uninformative teratologies in which the developmental system had been given a nonspecific jolt, making it veer off in a peculiar direction. They were unruly freaks of nature that would not repay further study. It has only been during the last 10 years, through the combined use of molecular and genetic approaches, that floral aberrations have started to regain the attention of biologists. My first scientific encounter with monstrous flowers occurred in 1983, when I was being interviewed for a position at the John Innes Institute, Norwich. During my visit, Brian Harrison and Rosemary Carpenter took me around the fields and glasshouses to show me some of their mutant Antirrhinum plants (snapdragons). Brian had been working on unstable Antirrhinum mutants with variegated flowers at the John Innes since the late 1950s, and was about to retire. Rosemary had joined Brian in the early 1960s and over the years they had built up an extraordinarily rich collection of genetic material. They delighted in showing me Antirrhinum flowers with all sorts of colours and patterns. I asked whether they had any examples of mutants with altered flower morphology. As if by magic, they produced two remarkable plants. One had small green flowers, with no obvious petals and a column of tissue projecting from the centre. They called this mutant 'coy' because of the rather modest appearance of its flowers (it turned out to be an allele of the globosa gene, described below). The second plant was even more dramatic: it had perfect radially symmetrical flowers instead of the normal bilaterally symmetrical type. What really struck me about these mutants was how such profound alterations in the basic plan of the flower could be produced without any apparent effect on the rest of the plant. The mutations were clearly affecting genes with very specific and fundamental roles in flower development. I had previously done a Ph.D on Drosophila with Gabby Dover, on the evolution and genetic behaviour of ribosomal DNA (Coen et al., 1982; Coen and Dover, 1983). Through this, I had become interested in applying genetic and molecular methods to the study of development and evolution. One problem that had caught my attention was the genetic control of floral development in Primula vulgaris. Individuals of this species are of two sexually incompatible types with distinct floral morphologies. These differences are controlled by a complex genetic locus and I was interested in trying to isolate and study this at the molecular level. I obtained a postdoctoral fellowship to work on the problem and Dick Flavell kindly agreed to let me carry out the work in his lab at the Plant Breeding Institute, Cambridge. After working on this for a while, it became clear to me that trying to develop the required molecular and genetic tools in the Primula system from scratch was over-ambitious. The Antirrhinum material seemed to offer a much better opportunity to follow up the interests I had now developed in plants. There was already a clear possibility of applying molecular analysis to the system because the group of Heinz Saedler and Hans Sommer at the Max-Planck Institute, Cologne had just shown that some of the Antirrhinum variegated flower mutants, sent to them by Brian Harrison, were caused by transposon insertions (Wienand et al., 1982; Bonas et al., 1984a,b). Transposon behaviour and the way it influenced gene expression was itself a fascinating problem that had now become open to molecular and genetic analysis. In the longer term, I felt that the availability of cloned transposons might also provide a way to study the flower developmental mutants that had so struck me. These were some of the main reasons for my wanting to pursue the Antirrhinum system in Norwich. There was, however, another important reason. Over the years, Rosemary Carpenter had acquired unrivalled experience in Antirrhinum genetics. She had systematically maintained records and genetic pedigrees going back 20 years and was very enthusiastic about continuing with her work and collaborating with others. Having had little experience in plant genetics myself, working with Rosemary seemed to hlve the makings of a perfect