My life with the Red Queen in fishery genetics

Abstract

I had the privilege of being a part of fishery genetics from its start in the 1970s,  when protein electrophoresis was first used to identify stocks of commercially important fishes and shellfishes. Research questions in fishery genetics have evolved tremendously over the past few decades, as new molecular techniques changed the nature of the questions that could be posed. The development of new DNA methods spurred the development of new theoretical models, new statistical methods and an unending stream of computer programs. These developments have led to conceptual shifts in the understanding of natural populations and how to manage them. Twists and turns in the history of fishery genetics parallel the story of the Red Queen in Lewis Carroll's Through the Looking Glass. I have had to “run in place” during my career to learn new techniques and ideas that, in turn, have been replaced with yet newer ones. However, my personal challenge throughout my career has been to develop “scientific intuition” to find underlying causes in datasets. Even though we now have the ability to sequence entire genomes, it is premature to announce the “end of history” for technical and conceptual advances in fishery genetics.

Introduction

This essay stems from my experiences in fishery genetics from its inception in the early 1970s and focuses on three threads running through its short history: a persistent turnover of molecular techniques, the development of computer software to implement new statistical methods and the evolution of the conceptual framework in population genetics. This history parallels Van Valen's (1977) Red Queen metaphor from Lewis Carroll's Through the Looking Glass (Carroll, 1872) to describe ever-changing physical and biological environments that force species to “run in place” to keep pace with environmental shifts. Species diversity and ecosystem functions remain largely unchanged while the cast of species changes. In a similar fashion, I have been in a perpetual state of running in place over my career to learn new molecular techniques and deal with an unending torrent of new computer programs. The Red Queen never stops.

The eighth row

Before embarking on a narrative about the conceptual evolution of fishery genetics, I would like to recognize the human dimension behind research activities that is not always apparent in a published article. The Red Queen is a chess piece in a large game being played out in “chess-board land,” and she advices Alice, who starts the story as a pawn, that if she can work her way to the eighth row of the chessboard, she too can become any piece she wants to be, even a queen. In a society, where fathers tell their children “you too can become the President,” I was imprinted with an imperative to climb the leadership ladder. But I was an introvert and did not have the silver tongue to give inspiring talks at conferences, nor could I catch the interest of job interviewers. At critical moments, I got an ice cream-like brain freeze.

However, there are many other paths to a successful career in science. My path toward the eighth row was sign-posted by my formal education, by mentors who passed on their excitement for research, and by colleagues and peer reviewers who provided comments that improved my ideas. More fundamentally, progress along a pathway was guided by my particular configuration of cognitive skills (Jung 1921; and his conceptual offspring). Many of us have taken personality “tests” administered by human resource managers with the intent of helping us appreciate work-place diversity. As it turns out, introversion isn't bad for writing manuscripts. However, beyond extroversion-introversion, other cognitive variables shape our choices of research problems and how we address them.

Allow me to digress further from the topic of fishery genetics to explore a philosophical model of knowledge that explains the source of my motivations and satisfactions during my career. The Renaissance philosopher Baruch Spinoza (1632–1677) outlined various kinds of knowledge. Science is all about understanding old knowledge and creating new concepts. One kind of knowledge comes from experience; for example, after I took a course in plant taxonomy, it was easy for me to recognize plants that belonged to the “rose” family because they had flowers parts in groups of five. A second form of knowledge arises from the use of reason, which has been used to establish numerous “laws of nature” and common notions of the world. In population genetics, a fundamental principle is the Hardy–Weinberg prediction of genotypic proportions in a randomly mating population. Deviations from these proportions often lead to insights into population dynamics.

Spinoza, however, thought that the highest form of knowledge came from intuition, which combines observation with reason, but adds scientific imagination, to arrive at an understanding of a problem. Intuitive “aha” moments are important, not only for solving research problems, but also for finding a story to communicate an insight. The creative use of metaphor, like van Valen's use of the Red Queen, is an earmark of effective scientific writing. In my career, it has been these intuitive dopamine-producing moments that have motivated me to work evenings and weekends on manuscripts throughout my career. My wife comments on my discipline, as I spend hours hunched over my laptop. “No,” I tell her, “I am a lab rat pressing a lever for another dopamine reward.”

The evolution of genetic applications in fishery management

Genetics is now a fundamental tool of harvest and hatchery management, particularly for the five species of salmon inhabiting the rivers and streams along western North America from California to the Bering Sea in Alaska (Myers et al., 1998). Genetic markers have been especially useful in the toolboxes of State and Federal resource managers for delineating populations of salmon, because of their homing behaviour to natal streams to spawn. The identification of genetically discrete populations, which are assumed to be largely adapted to local environmental conditions, is a starting point for harvest management, for the protection of genetic diversity and the conservation of endangered populations (Waples, 1991). Many salmon populations are heavily supplemented with hatchery-raised fish, and the straying of these fish can present ecological and genetic problems for wild populations when strays mix with wild fish in natural spawning areas (Grant, 2012).

Molecular markers, however, have been less useful for managing marine species, especially pelagic spawners, because of the large potential for dispersal between areas. Hence, the focus of genetic research in the marine realm has been on broad-scale evolutionary issues, including phylogeny, biogeography, and taxonomy. As with salmonids, the aquaculture of marine fishes, invertebrates and seaweeds also poses a potential threat to wild populations that requires attention to brood stock origins, captive population management and effects of hatchery strays (Grant et al., 2017). My research activities have wandered into each of these areas, as funding and life's opportunities presented themselves.

Molecular methods

My first encounter with molecular methods was as an ecologist at the genetics lab at the NOAA Northwest Fisheries (Science) Center in Seattle. I had just started PhD studies in the Department of Botany at the University of Washington with a focus on seaweeds (Grant and Horner, 1976; Grant, 1977) and seagrass (Phillips et al., 1983) ecology. I heard about a course being taught on a new technique in population genetics involving the electrophoretic analysis of proteins, chiefly enzymes. Fred Utter's crowded lab at NOAA bristled with excitement, as a diverse group of students set about studying salmon, trout, invertebrates and plants. The results of this work were quickly incorporated into fishery management (Allendorf and Utter, 1979). My attempts to extract live enzymes from seaweeds were unsuccessful, so I settled on a study of a common intertidal whelk to satisfy the requirements of the course (Grant and Utter, 1988). Animals bled enzymes much better than seaweeds.

A successful project on whelk genetics led to an offer to take on a project funded by the Alaska Department of Fish and Game to identify the population genetic components of salmon in a mixed-stock fishery in Cook Inlet. The identification of stocks, as salmon returned to natal lakes and streams to spawn, was an essential part of fishery management. A minimal number of spawners had to “escape” the gauntlet of fishing boats in the Inlet to allow adequate numbers of spawners. Traditionally, stream-side counts, aerial surveys and fish scales were the chief tools for managing salmon harvests. With the help of George Milner, another grad student in the lab who had written a FORTRAN program in the late 1970s to sort out mixed stocks, we were able to estimate the natal origins of the sockeye salmon returning to Cook Inlet rivers (Grant et al., 1980). These primitive estimates were consistent with migration patterns identified two decades later (Seeb et al., 2000), indicating stability in migration pathways to natal streams and rivers.

A large, heavily marked, comprehensive diagram of metabolic pathways was mounted on the wall of the lab and was frequently consulted to troubleshoot problems with enzyme stains and to develop new staining cocktails. A pinch of arsenic perhaps was needed to preserve the product of an enzyme reaction, or more NAD⁺ cofactor was needed to help a reaction along. Recipes were approximate, and the components of live-enzyme stains were measured in “match-head” units or drops of liquid (Aebersold et al., 1987). The electrophoretic gadgets we used were home-made, consisting of a slab of double-pane glass, plexiglass strips held in place with large binder clips, kitchen paraphernalia and platinum-tipped wires leading from a DC power supply. The lab had 22 power units and all were in use on most days. The potato starch we used as a supporting matrix to separate allozymes in an electric field sometimes had bits of potato skin in it; but it did not seem to matter. These simple, but effective, procedures were the distant ancestors of the complex and costly infrastructure needed to generate DNA data today.

A critical step in gathering genotypic data was the interpretation of complex banding patterns on a starch gel. Proteins came in a variety of quaternary structures, which influenced how homo- and heterozygotes appeared on a gel. An early focus in developing genetic methods was to demonstrate through breeding experiments that the electrophoretic variants followed the rules of Mendelian inheritance (May et al., 1975, among numerous other publications). At that time, the basement of the NOAA lab had a running water system, which was used to incubate batches of salmon eggs from parents with known genotypes. In a matter of weeks, eggs developed into larvae large enough to yield crude protein extracts that could be tested with electrophoresis for Mendelian ratios. However, for most other species, breeding experiments were not possible, so the Mendelian nature of the allozymes had to be inferred from banding patterns and from the fit to Hardy–Weinberg genotype frequency expectations.

An early genetic study was on Pacific herring populations across the North Pacific. Getting fresh tissue samples from far off places was a major challenge. But with the help of several herring biologists, I was able to amass 31 samples from across the species’ geographic range from Korea to San Francisco Bay. Allozymes showed an unexpected sharp genetic discontinuity between the Asian and North American sides of the North Pacific that had to be a legacy of ice-age glaciations (Grant and Utter, 1984). The results of this study sparked a career-long interest in marine biogeography and, in particular, in the effects that Pleistocene glaciations had on marine fishes (Grant and Utter, 1980; Grant et al., 1984b, Canino et al., 2010), invertebrates (Grant et al., 2014) and kelps (Grant and Bringloe, 2020; Grant and Chenoweth, 2021) around the North Pacific.

In the early 1980s,  I took a post-doc position in Cape Town. I thought I'd stay a year or so to see the wildlife of southern Africa, but the stay extended 15 years and ended with a professorship at the University of the Witwatersrand (“Wits,” pronounced “Vits,” as the name has Afrikaans origins) in Johannesburg. On landing in Cape Town, my initial project, as part of the large Benguela Ecology Programme, was to study the genetic stock structure of southern African anchovies (Engraulis) and sardines (Sardinops) and, eventually, hakes (Merluccius). Allozymes revealed no genetic differences among anchovies or sardines along the shores of southern Africa (Grant, 1985; Grant and Leslie, 1996), but an interesting shift in allele frequencies occurred between Namibia and South Africa for one of the species of hake (Grant et al., 1987). This break was subsequently confirmed with microsatellite DNA (Henriques et al., 2016). This was a rare instance in the marine world in which genetic stock boundaries coincided with political boundaries.

I used every opportunity to collect samples of related species or populations around the globe. When colleagues went to an international conference, I set them up with a small cooler and gel packs that could be frozen, so they could bring back anchovies, sardines, anglerfish and hakes from a local fish market. These efforts led to insights into the biogeography and evolution of global populations of anchovies (Grant and Bowen, 1998; Grant et al., 2005), sardines (Parrish et al., 1989; Bowen and Grant, 1997), hakes (Grant and Leslie, 2001), and anglerfish (Grant and Leslie, 1993).

Student projects also led to exciting discoveries. One was a study of a “surfing” gastropod that moved quickly in the wash of sandy beach waves by extending its foot into a large “sail.” This whelk lacked a pelagic larval stage; it incubated eggs on its foot that were released as crawl-away juveniles that had larval shell sculpting typical of non-pelagic whelks. Thus, we expected a subdivided genetic population structure but found no allozyme frequency differences over 2500 kilometers (Grant and da Silva-Tatley, 1997). This brought into question the use of shell morphology in fossil gastropod shells to infer ancient genetic population structure. In another project on black mussels, the first allozyme gel showed the presence of a cryptic species, which we identified as the Mediterranean mussel (Grant et al., 1984a; Grant and Cherry, 1985). We caught the early stages of the invasion of this species, which in the decades since its introduction has re-organized rocky intertidal communities (Robinson et al., 2007).

The Red Queen prodded me again to learn new lab techniques. Protein electrophoresis detected variability only in gene products and revealed only a fraction of the variability in the underlying DNA. Like the shadows in Plato's cave, allozymes provided only a crude first glimpse of genomic variability. The restriction-enzyme fragment analyses of mitochondrial (mt) DNA appeared in the late 1970s and blossomed in the 1980s and 1990s. The analysis of maternally inherited, un-recombining mtDNA not only provided insights into patterns of diversity within and between populations, but more importantly yielded gene genealogies that revealed historical dispersals and colonizations (Avise et al., 1987). These methods spawned a vibrant discipline, “phylogeography,” conceptually located between phylogenetics and population genetics.

An early method of mtDNA analysis employed a class of enzymes, restriction enzymes isolated from bacteria, that produced DNA restriction fragment length polymorphisms (RFLP) (Smith and Wilcox, 1970). These fragments could be separated on an electrophoretic agar gel and visualized with ethidium bromide under a UV light. RFLP analysis of mtDNA provided estimates of genetic diversity but also insights into the histories of populations in a species or group of closely related species. Tissue samples had to be kept deep frozen in liquid nitrogen or –80°C freezers to keep the circular mtDNA intact so it could be isolated in ultracentrifuges that ran at 40 000 + RPMs for 40 hours. The faint band of mtDNA was carefully extracted from the ultracentrifuge tube, labelled with a radioactive tag such as ³²P, and run on an agarose gel. If all went well, a sheet of X-ray film picked up the labelled DNA fragments. Needless to say, early studies required considerable infrastructure that only a few labs had and produced datasets that were necessarily small. Even so, the method led to new insights into two species of southern African hakes (Becker et al., 1988).

Techniques lurched forward again. The development of the polymerase chain reaction (PCR; Mullis et al., 1986) and Sanger dideoxy DNA sequencing (Sanger and Coulson, 1975) were great strides forward to facilitate the analysis of mtDNA. Deep-frozen tissue samples were no longer required to isolate mtDNA. The availability of Sanger sequencing at modest costs allowed much larger sample sizes and led to a booming cottage industry of phylogeographic studies. DNA sequencing revealed mutations along the entire stretch of DNA, not just within restriction-enzyme recognition sites. In the early 1990s,  I spent a sabbatical in Brian Bowen's lab at the University of Florida to absorb this technology and to analyze the global samples of anchovies and sardines that I had cobbled together.

This collaboration was most fruitful. Mitochondrial DNA sequences showed that the five putative species of Indo-Pacific sardines (Sardinops) were populations of a single species. The close genetic relationships implied recent long-distance stepping-stone dispersals from southern Africa around the rim of the Pacific Ocean to Australia–New Zealand, Japan, California and Peru-Chile (Parish et al., 1989; Bowen and Grant, 1997; Grant and Bowen, 1998). Both allozymes and mtDNA confirmed that European sardines (Sardina) belonged in a separate genus.

We found a contrasting global phylogeographic pattern for anchovies. Unlike sardines, allozymes, and mtDNA provided evidence for four anchovy species. Three New World species occurred in California, Peru-Chile and Argentina, but a fourth, Old World species, included populations in Europe, southern Africa, Australia-New Zealand and Japan. As with sardines, a single anchovy species was widely distributed, but in a different global pattern. The oldest of the Old-World anchovy populations was in European waters. This population was connected to other populations with stepping stone dispersals via southern Africa to Australia-New Zealand and Japan (Grant and Bowen, 1998). While southern Africa may have been one of the stepping stones for dispersal across the Indian Ocean, the southern African population had low levels of mtDNA diversity and had haplotypes that were more closely related to haplotypes in European waters than to other haplotypes in southern Africa. This clearly indicated that the southern African population had gone extinct and had recently been recolonized by European anchovies (Grant and Bowen, 2006).

Mitochondrial DNA provided insights into the phylogeographic structures of populations and closely related species, but its usefulness to answer some questions was limited, because it represented a single gene marker of female population history. Nonetheless, phylogeographic methods still have a role to play as a rapid means of defining research problems and as a useful tool in developing countries (Bowen et al., 2014). The use of multiple markers provided more accurate and statistically more powerful inferences about population structure. Short repetitive DNA interspersed between coding regions in nuclear DNA, called microsatellites, opened an alternative window onto a population's historical and contemporary dynamics. Microsatellite DNA has large mutation rates that can resolve population events on short time scales. Since microsatellites are passed from one generation to the next by both parents, and not just one parent as mtDNA, they could be used to test several hypotheses of mating behaviour as well as population structure. For example, the analysis of organellar mtDNA and chloroplast DNA, along with microsatellites, detected hybrid individuals between incipient species of winged kelp in the Gulf of Alaska (Grant and Bringloe, 2020).

Over the past two decades, molecular methods lurched forward again with the development of methods that interrogated entire genomes, including RAD Sequencing (Etter et al., 2011), expressed sequence tags (EST, Bowman et al., 2011), whole genome sequencing (Johansen et al., 2009), and transcriptome analysis (Johansen et al., 2011; Montes et al., 2016). From a fishery management perspective, genomic surveys yield single nucleotide polymorphisms (SNPs) that can serve as diagnostic population markers. These markers serve to define biological population units that are at the heart of genetic resource management. Several methods have been developed to assay SNPs rapidly to shorten laboratory turnaround times and to provide real-time mixed-stock information for salmon fishery managers (Dann et al., 2013).

The next challenge for fishery genetics was to use these genomic methods to understand the adaptive landscape of wild populations. A core assumption, not only in the use of genetic markers to define stock boundaries, but also in stock assessments for harvest management and in the planning of marine protected areas, was that populations are uniquely adapted to local environments. Genomic methods now allow the identification of outlier genes and gene complexes that underlie adaptation (Petrou et al., 2021). Fishery biologists will have to decide which adaptations need protecting (Grant, 2012). This will be especially challenging in the coming decades when global climate warming is expected to massively alter local environments.

Statistical methods followed the development of new kinds of genetic data

Advances in molecular methods routinely required the development of new statistical approaches. At the start of the allozyme boom, we used the university's mainframe computer, tediously punching holes into computer cards with noisy key punch machines. We submitted our data-cards, along with FORTRAN instructions on what to do with the data, at the intake window and waited, sometimes an hour or more to see our results, only to find an error buried in our deck of cards. Sometimes, long waits reflected large amounts of CPUs required by the computer program. Fellow grad student George Milner had written a mixed-stock computer program and produced an iterative maximum likelihood algorithm that took a lot of computer time, even for modest datasets. He once got a notice that whatever he was doing took 25% of the university's mainframe capacity for several hours. His efforts yielded one of the first programs to predict the destinations of salmon returning to natal streams to spawn (Grant et al., 1980). Mixed-stock analysis is now a cornerstone of harvest management and conservation of salmon (Hess et al., 2016; Beacham et al., 2020; Bradbury et al., 2021), and migratory marine species (Bowen et al., 2007).

Computer programs to analyze data were scarce in the early days of fishery genetics, and most population geneticists had to learn a computer programming language. We continually had to modify our programs as new statistics came along. An important skill was to decipher runtime error messages. The problem was not usually at the step in the code with the error flag, but earlier in the program, perhaps a miscalculated variable that caused a hiccup downstream. I played with Pascal and True Basic, before programming shifted to complex computer languages, such as C and C++, and left me behind. The use of ready-made programs that produce “black-box” output, however, can be problematic because ignoring the underlying assumptions in the program can lead to misinterpretations of data (Grant, 2015; Grant et al., 2016). To address this problem, the use of programs written in the now-popular “R” interpretive language provides transparency in the methodological guts of a program.

When desktop computers appeared in the early 1980s,  I spent 40% of my postdoc salary on a floppy-disk desktop computer that gave me a giant boost, not only with data analyses, but also with writing. By the mid-1980s,  freely available computer packages offered a tremendous range of data analyses. Programming specialists have taken over the task of producing data-analysis software, so today population geneticists only have to learn how to implement the various programs. One downside, however, is that many researchers do not pay attention to the critical assumptions used in a program (Karl et al., 2012; Grant, 2015; Grant et al., 2016).

The early analyses of genetic data were made largely within a “frequentist” framework (Swofford and Sealander, 1981) before personal computer speed and memory allowed the development of resampling and Markov chain Monte Carlo (MCMC) methods to test hypotheses with likelihood and Bayesian approaches. Today, an incredible number of programs can be freely downloaded, and the number increases monthly to accommodate the large data sets being generated with genomic methods and to implement new theoretical models.

Conceptual shifts in fishery genetics

When I walked into Fred Utter's lab at NOAA just across the Lake Union-Lake Washington canal from the University of Washington in 1975 to learn a new molecular technique, I entered a conceptual world shaped by the 1930s concept of genes-on-a-string proposed by Beadle and Tatum (1941) and by the population genetic equations of RA Fisher, JSB Haldane and S Wright (FHW). In a series of theoretical papers, FHW demonstrated that microevolutionary changes within and among populations could be explained by breeding behaviour, random genetic drift, and gene flow, or by natural selection. This theoretical framework continued to grow, but applications to natural populations were lacking until protein electrophoresis came on the scene.

Genetic studies of wild populations were largely limited to morphological traits and chromosomal polymorphisms that were tediously scored with a microscope (e.g. Dobzhansky, 1958). Beginning in the 1930s,  Dobzhansky used this framework to interpret his observations of chromosomal variants and inversions in fruit flies. His studies showed that wild populations contained considerable genetic variability. This was contrary to the view at that time that species converged on a single “wild-type” genotype. The promise of the FHW framework to understand population dynamics was fulfilled only with the advent of molecular techniques, beginning with protein electrophoresis and continuing with the development of DNA markers, and now with genomic methods.

Assumption of selective neutrality

A major problem in the application of electrophoretic data to fishery management was to gauge the role that natural selection played in shaping allele-frequency differences among populations. The most useful FHW models in fishery genetics assumed that natural selection was unimportant, even though the selection was the buzz word of evolutionary biology. The solution came with the work of Kimura (1968), who postulated that most evolutionary changes, including electrophoretic variants, were neutral to the effects of natural selection. Most, but not all, population geneticists grabbed onto this idea like a drowning man clutching a life-saver ring. Fishery geneticists followed suit and used FHW models to produce estimates of genetic diversity, inbreeding and gene flow. The use of allozymes markers was often promoted to funding agencies for its ability to provide estimates of “connectivity” among populations. However, the evolutionary time-scales estimates of gene flow are vastly larger than the ecological time-scales used by fishery harvest managers. Genetics could not replace the insights that tagging data and decadal times-series of stock assessments provided.

Not everyone was convinced that electrophoretic alleles were neutral to natural selection (Crawford and Powers, 1989), and a bitter debate in the population genetics community spilled over into fishery genetics. This debate played out with the use of electrophoretic data to manage fishery populations, primarily salmon populations. Protein electrophoresis was ideally suited to describing the genetic structures of salmon populations, because populations were isolated by faithful homing to natal spawn sites and because salmon populations were small enough to experience random drift that produced differences between populations. The bulk of not only allozyme markers, but later mtDNA, microsatellite and single nucleotide polymorphism (SNPs) markers were assumed to be neutral to selection to estimate population parameters thought to be useful to fishery managers. The most useful applications of genetic data lay in estimating population genetic diversities, delineating population boundaries and estimating population components in mixed-stock fisheries. The development of large multilocus profiles has extended this approach to identify parent-offspring relationships in a small population (Steele et al., 2019).

However, the Red Queen keeps us running. New state-of-the-art technologies now allow us to survey DNA variability throughout an entire genome and to experimentally understand the effects of environmental drivers on regulating genetic variability. The molecular analysis of the genomes in model organisms is beginning to show the large extent that transcription factors, DNA methylation and re-arrangements of DNA sequences influence morphology and behaviour in a growing discipline called “epigenetics.” There is no evidence that modified DNA can mimic beads-on-a-string inheritance. Epigenetic changes may explain the rapid domestication of salmon in hatcheries, for example (Grant, 2012; Grant et al., 2017). Unfortunately, I am retiring at an exciting time in the history of genetics. New laboratory methods and a burgeoning field of bioinformatics are now producing insights into the adaptive responses of invertebrates and fishes to environmental change. Natural selection is back, and in a big way!

Adaptive variability

Assumptions of selective neutrality of molecular markers have taken a back seat to the complexities that genomic methods are revealing. Whole genome sequencing and transcriptome sequencing provide windows into the dynamics of adaptation and evolutionary dynamics. For example, where allozymes (Grant and Utter, 1984) and mtDNA (Grant et al., 2012; Liu et al., 2012) showed only small genetic differences between populations of Pacific herring in the Northeast Pacific Ocean, a genomic survey of populations revealed variability in genes controlling spawn timing (Petrou et al., 2021).

The genomic revolution has just begun and will undoubtedly provide a level of information to fisheries managers and conservationists that has not previously been available. This will create a challenge for resource managers in how to use small-scale adaptive variation in a particular management approach. Current studies show that adaptation occurs on much finer spatial scales than stock structure estimated with “neutral” genetic markers. With these new insights, a challenge to fishery managers will be to balance both economic and biological concerns. Another challenge will be to understand the temporal dynamics of adaptation. Populations are continually expanding and contracting, and even perishing on decadal time scales. The use of genetics in fishery management is at the junction of economics, contemporary ecology and evolutionary biology, and there are no road maps to consult. Can adaptive change keep pace with human exploitation and a changing environment?

The Red Queen is alive and well

In a broader perspective, my experiences as a researcher have followed just one of an infinite number of conceptual pathways. This path was sign-posted by my education, in and out of the classroom, and by my particular cognitive skills and personality traits. I've irritated some colleagues by writing overly detailed manuscripts—introverted thinking—and by pursing ideas that seemed irrelevant to job assignments, particularly in fishery management agencies with narrowly define mission statements—curiosity. The eighth row for me in the conceptual landscape of fishery genetics, the equivalent of the Red Queen's chessboard-land, was not a position of up front leadership, but the many exciting moments of creative insight that appeared as I wrote.

To illustrate how our personality influences our approach to a problem, the organizers of a workshop in Portugal on marine biogeography of the Atlantic Ocean recorded interviews of speakers to pass on advice to young researchers. One of the questions was “what advice can you give about finding research questions.” I answered that it was important to spend time in a library reading about the thoughts of others and giving yourself quiet time to synthesize what you had read. Cliff Cunningham, on the other hand, said that it was important to attend conferences and workshops to hear others talk about their research. These different approaches clearly reflect personality differences. As with biological evolution, we respond in different ways to the ever-changing landscape of techniques and concepts. The Red Queen beckons us on.

Data availability

There is no data associated with the article.

Notes

Food for Thought articles are essays in which the author provides their perspective on a research area, topic, or issue. They are intended to provide contributors with a forum through which to air their own views and experiences, with few of the constraints that govern standard research articles. This Food for Thought article is one in a series solicited from leading figures in the fisheries and aquatic sciences community. The objective is to offer lessons and insights from their careers in an accessible and pedagogical form from which the community, and particularly early career scientists, will benefit.

The International Council for the Exploration of the Sea (ICES) and Oxford University Press are pleased to make these Food for Thought articles immediately available as free access documents.

Acknowledgements

Several individuals have played pivotal roles in guiding my career by providing job opportunities. Ron Phillips (seagrass ecology) and Bob Vadas (seaweed ecology) gave me a front seat as an undergraduate in how biologists think during field trips to research diving sites in Puget Sound. Fred Utter not only introduced me to the excitement of population genetics, but also connected me with research project on salmon in Alaska that led to the development of new management techniques. I have also had the privilege of working with George Milner, Brian Bowen, and Rob Leslie on several projects that caught all of our creative imaginations. I am also grateful for the support of numerous laboratory support staff at NOAA and the Alaska Department of Fish and Game. Finally, Brian Bowen and my wife, Marianne, provided a sounding board for this essay. I am fortunate to have a wife whose understanding of the intricacies of English, as her second language, surpasses my own.

References