Abstract
In the later part of the 1980s, the time was ripe for identifying genes controlling flower development. In that pregenomic era, the easiest way to do this was to induce random mutations in seeds by chemical mutagens (or irradiation) and to screen thousands of plants for those with phenotypes specifically defective in floral morphogenesis. Here, we discuss the results of premolecular screens for flower development mutants in Arabidopsis thaliana, carried out at Caltech and Monash University, emphasizing the usefulness of saturation mutagenesis, multiple alleles to identify full loss-of-function, conclusions based on multiple mutant analyses, and from screens for enhancer and suppressor modifiers of original mutant phenotypes. One outcome was a series of mutants that led to the ABC floral organ identity model (AP1, AP2, AP3, PI, and AG). In addition, genes controlling flower meristem identity (AP1, CAL, and LFY), floral meristem size (CLV1 and CLV3), development of individual floral organ types (CRC, SPT, and PTL), and inflorescence meristem properties (TFL1, PIN1, and PID) were defined. These occurrences formed targets for cloning that eventually helped lead to an understanding of transcriptional control of the identity of floral organs and flower meristems, signaling within meristems, and the role of auxin in initiating floral organogenesis. These findings in Arabidopsis are now being applied to investigate how orthologous and paralogous genes act in other flowering plants, allowing us to wander in the fertile fields of evo-devo.