As someone that has worked on fly behavior for many years, I’ve become more and more fascinated with the beginnings of the field. It always amazes me how serendipitous science is, and how outside influences such as what is going on in the world, who your colleagues are, and family life, can influence scientific decisions and events. A lot of what I’ve learned about the origins of the fly behavior field are from others that have been in the field a lot longer than I have.
Shortly after Semour Benzer died, Ralph Greenspan wrote a really interesting perspective in Current Biology called “The origins of behavioral genetics” (2008). From this perspective, I learned that the first person to study flies in the lab wasn’t Thomas Hunt Morgan (as I had naively thought) but William E. Castle, who followed the suggestion by entomologist colleague C. W. Woodsworth to study fruit flies in the lab in 1901. Castle’s student F. W. Carpenter worked on phototaxis, geotaxis and mechanosensation (Carpenter, 1905), and Castle and his students later published a study on fertility (Castle et al, 1906) and olfactory responses (Barrows, 1907).
Morgan’s FlyRoom was the site of the first Drosophila behavior genetics experiments. A.H. Sturtevant published on genetic variation in courtship behavior (Sturtevant, 1915) and R.S. McEwan studied phototropic and geotropic responses (McEwan, 1918; McEwan, 1925).
The next phase of inquiry brought the emergence of quantitative genetics, and the question of whether a single gene could influence behavior. J.P. Scott recognized that the behavioral differences in white or brown mutants might be due to genetic backgrounds independent of these genes (Scott, 1943). Jerry Hirsch was a real pioneer in the quantitative genetics to the study of behavior and introduced chromosome analysis to map behavior traits like geotaxis (Erlenmeyer-Kimling & Hirsch, 1961). The consistent conclusions that emerged from Hirsch’s studies was that behavior has a complex and mulitgenic architecture.
That left some contradictions in the literature though. If behavioral traits were multigenic, how did single genes like white and brown affect phototaxis (Brown & Hall, 1936), Barr affect optomotor response (Hecht & Wald, 1934; Kalmus, 1943), and ebony affect mating (Rendel, 1951)? A more ethological view of the genetic origins of behavior emerged looking at courtship (Bastock, 1956); Bastock & Manning, 1955), aggression and mating success (Dow & Von Schilcher, 1975).
In the 1970’s and 1980’s, the mutagenesis approach to understand mechanisms underlying Drosophila behavior (now termed Drosophila neurogenetics) was largely spearheaded by Seymour Benzer and his trainees. Work in his lab was key for identifying that single genes influence behavior and revealing the neural mechanisms through which this could occur (Benzer, 1973). Topics in his lab included the famous genetic dissection of circadian rhythms, learning and memory, neurodevelopment, neurodegeneration, cell fate, feeding, pain, and many more. Some of the most interesting fly mutant names came out of this period (dunce, painless, drop-dead, sevenless, period, etc). For more information on this, I recommend reading Time, Love, Memory by Jonathan Weiner and the many interesting biographical write-ups that were published after his death in 2007 (Bonini 2008; Greenspan 2008; Quinn 2008; Dudai 2008; Tanouye 2008; Anderson & Brenner 2008; Jan & Jan 2008).
These combined complex trait and single gene approaches are now recognized being complementary, and helped scientists converge on the idea of multifactor inheritance in which ‘major genes’ can have large effects on behavior and ‘minor genes’ have individually small effects (Roderick & Schlager, Chapter 9). Later Drosophila studies demonstrated this principle occurs in nature through some of the best examples of how natural variation in a single (major) gene influences behavior. Natural variants in the Alcohol dehydrogenase (Adh) gene have differences in alcohol metabolism and associated behaviors (McDonald & Ayala, 1978; Sampsell, 1977; Watanabe & Watanabe, 1977). The foraging gene encodes a cGMP-dependent protein kinase and has two variants that differ in their foraging and feeding behaviors: rovers and sitter (Osborne et al, 1997). The clock gene has a long form and a short form that vary in their circadian rhythms along geographical clines (Sawyer et al, 1997).
The take home message here is that fly behavior has been part of the Drosophila field since its very inception. Perhaps that shouldn't surprise us since flies behavior is amazing to watch, and in a time when microscopes weren't near as powerful as they are now, this would have been one of the most convenient ways of obtaining repeatable phenotypes.
It’s amazing to see how far we have come since these early days. It is impressive to see how technology has driven how we think about the ways genes influence behavior. From genetic sequencing to single-cell multi-ome sequencing, from the GAL4>UAS system to the split-Gal4 and connectome approaches, and from watching flies behaving in vials and Petri dishes to using computer vision and machine learning to measure fly behavior, we are now at a time where we can understand how the environment and genes interact to influence subtle changes in behavior at a nearly single cell level. I’m eager to see how this plays out in our functional understanding of complex behaviors such as memory and addiction.