A crew of patches

Given the present obsession with pattern formation, it is hard to appreciate that 30 years ago there were but few eccentrics who knew what the words meant. In those days, any who tried to understand patterns searched throughout zoology and botany for colleagues who shared their interest. Topics of research included the spacing of hairs on sheep or of stomata in plants, and the orientation of insect scales and bristles. Retinotectal projections were also in vogue as they seemed to incorporate several major problems at once: growth, positional information, polarity and neuronal specificity. Having every problem to cope with at once was seen as an advantage! Some molecular biology meetings had a menagerie day — invariably the last day — in which, together with a talk on retinotectal projections, oddballs like myself or Lewis Wolpert were wheeled on as a reminder that all was not yet solved. This kind of thing does not happen much nowadays; my impression is that even scientists in the field of a meeting can be excluded just because they have unconventional views. The decision to devote the rest of my research life to understanding pattern formation was like a conversion; I was on a Loftleider plane returning from a post-doc in the USA in 1967, and the conviction came to me about an hour out of New York. While in the States I had been messing about with several problems, but I had become certain that genetics was the best way forward for the study of development. Therefore, I was en route to the genetics department in Cambridge, where I had been offered the chance to mess about a bit more without anybody telling me what to do or, if any of it was publishable, putting their name on my paper. Happy days! Wedded to the experimental opportunities offered by various hemimetabolic insects such as the milkweed bug, Oncopeltus, I had been transplanting pieces of cuticle to study aspects of pattern — such as how hormones influence bristle development. Having decided to make mutations to mark my grafts I tried X-rays, with some success. I noticed, however, that the bugs, which have orange epidermal cells growing on a transparent cuticle, grew up with spots of different colours, such as red, yellow, white and transparent. Not knowing which age would be best for mutagenesis I tried irradiating them at different stages and noticed that the earlier the irradiation the bigger the patches of coloured cells. At this point (late 1969) I was offered a junior post at the MRC Laboratory of Molecular Biology, as Sydney Brenner and Francis Crick had decided to collect people with different scientific backgrounds (me being the bug man — or ‘gogga man’ according to Brenner, a South African) and give them a free rein to try and get at developmental genetics. It was sometime around then that I found a paper by Peter Bryant and Howard Schneiderman [1] on the clonal analysis of the development of the Drosophila leg. (Unknown to me, Antonio GarciaBellido had used similar methods to study the Drosophila wing.) As a side project (I was also beginning a collaboration with Crick on gradients in pattern formation) I began to follow up the clonal analysis described by Bryant and Schneiderman. They had pointed out that from the size of the clones as a proportion of the whole, one could estimate the number of cells in the primordia at the time of irradiation, and from the rate of change of clone size, one could follow the division rate. They also argued that if a clone crossed from one structure to another (such as tibia to femur) it had to mean that, at the time the clone was induced, the mother cell was not determined as to its fate with respect to the two structures. They argued that when the structure that will form the leg (the leg disc) is first set aside it contains several tens of cells and that, in general, determination is a process that affects groups rather than single cells. I was impressed with the paper and the logic, and it was a great help that I could only find a few papers on pattern formation each year, so I gave it plenty of attention. I applied their logic to my coloured patches and looked at the abdomen of Oncopeltus. Insect segments had not been defined, other than by anatomical criteria, and I soon noticed that if I irradiated the early embryo, clones — for that is what they were — crossed from one segment to another. After a certain developmental stage, however, my clones always respected lines between segments (Figure 1). I realised that this meant that at some time segments were determined, and I crudely tried to estimate the number of primordial cells. The results defined segments as units of cell lineage and delineated their boundaries; such lineage units were later called compartments by Magazine R71