... it turned out not to be a happy experience. I didn't really fit in at the Academy. I realised that my father was right when he told me that most things in life 'tell' better than they 'live'. So although I had a lot of 'sea stories' to tell, I didn't always like being there. Sailing itself was mostly very boring.

The professors who interviewed me were amused by my transcript. They saw that I had courses like knot tying and marine seamanship. Nevertheless, they admitted me on probation, and I had to make up much of an entire undergraduate education during the first year.

I learned how to design reactors, which I didn't find much more interesting than sailing on ships.

... then when I got there, and they saw that I had a Ph.D., it was too late for them to say no.

... the singular event in my professional life happened: I found a mentor, Aharon Katchalsky, who was a visiting professor at Berkeley at the time and was looking for a student who knew something about thermodynamics. I had taught thermodynamics at Columbia during graduate school, and that gave me the hubris to approach Katchalsky, who was recognised as one of the great experts in the field. He must have seen through my pretences immediately, but he was a kind man and, for reasons I'll never really understand, he took me back to the Weizmann Institute as a postdoc. That was the turning point in my professional life. I believe that, for most scientists, a mentor is nearly indispensable; it surely was for me, and I could not have imagined a better one than Katchalsky. All I ever became as a scientist I owe to his influence and inspiration. Tragically, shortly after my postdoctoral stint, he was murdered at the Tel Aviv airport by terrorists as he was returning to become President of Israel.

This paper describes the development of a generalised topological representation for thermodynamic systems. The elements and topology of the representation provide a conceptual reticulation of the real system into elements drawn from a set of ideal, but physically realisable, lumped-parameter components. The representation is in the form of bond-graph notation.

The thrust of modern biology is toward the investigation of increasingly complex structures. It has become almost a cliché to acknowledge that living entities are extremely complicated, heterogeneous, non-linear systems. They are based on a subtle interplay between the energetic rate processes of transport, reaction and conformational change, on the one hand, and cybernetic flows of information, whose regulatory effects are not proportional to their energetic level, on the other. The theory outlined in this paper is designed specifically for the treatment of coupled, non-linear, time-dependent thermodynamic processes in heterogeneous media and is also capable of incorporating non-energetic, informational flows. The general formalism employs a simple and intuitive graphical notation which emphasises the topological relations of the system under consideration.

A biological population can be viewed as a distributed parameter dynamical system. The dynamics of host-parasite systems are investigated by coupling populations via age specific interactions; analysis and simulation results are compared with experiments performed on a laboratory ecosystem. A number of novel dynamical effects emerge from the model which have some interesting ecological consequences.

The entomology department at Berkeley offered Simon Levin a position as their resident mathematical population biologist and, when he turned it down, he suggested they ask me. They hired me, and so I became a population biologist. Moving to a biology department turned out to be the best professional decision I ever made. It forced me to teach biology students and absorb the culture of biologists that values experiments and solving particular problems far more than general theories. In biology, I've learned, it's easier to make a theory of everything than a theory of something.

The summer after taking the job in population biology I met Ed Wilson in a bar at Woods Hole, and that led to a book on ant castes. While working with Ed, I met Bob May at a Gordon Conference, and that led to our work on population chaos. Later, at another Gordon Conference, I heard Beth Burnside lecture on embryology and that evening I met Garry Odell. Together we started working on a model for gastrulation, and that launched my morphogenesis period. About then my new graduate student, Pere Alberch, saw the connections between our morphogenesis models and evolutionary trends, and that led to my work in 'evo-devo'. Then I met Jim Murray, who invited me to Oxford for a sabbatical because he was interested in morphogenesis too. That led to more developmental biology modelling and pattern formation, and to cell motility. From there it was but a short jump to modelling molecular motors. A chance meeting with Dale Kaiser at a conference led me into modelling bacterial locomotion and pattern formation. It seems that, as I grew older, my horizons contracted to smaller things rather than, as is more typical, expanded to larger issues.

Many biological populations breed seasonally and have nonoverlapping generations, so that their dynamics are described by first-order difference equations, $N_{t+1} = F(N_{t})$. In many cases, $F(N)$ as a function of $N$ will have a hump. We show, very generally, that as such a hump steepens, the dynamics goes from a stable point, to a bifurcating hierarchy of stable cycles of period $2^{n}$, into a region of chaotic behaviour where the population exhibits an apparently random sequence of "outbreaks" followed by "crashes." We give a detailed account of the underlying mathematics of this process and review other situations (in two- and higher dimensional systems, or in differential equation systems) where apparently random dynamics can arise from bifurcation processes. This complicated behaviour, in simple deterministic models, can have disturbing implications for the analysis and interpretation of biological data.

... it could be argued that a study of very simple nonlinear difference equations ... should be part of high school or elementary college mathematics courses. They would enrich the intuition of students who are currently nurtured on a diet of almost exclusively linear problems.

Most schemes for embryonic pattern formation are built around the notion of lateral inhibition. Models of this type arise in many settings, and all share some common characteristics. In this paper we examine a number of pattern formation models and show how the phenomenon of lateral inhibition constrains the possible geometries that can arise.

Oster's intervention into the assumed divide between mathematical physics and biology epitomised his creative and cross-disciplinary approach to scientific research. Oster also developed mathematical models to quantitatively explore biological mechanisms, transforming scientists' understanding of biomolecules and their systems.

George Oster often said that people are what is most important in science. He always pulled together a very interdisciplinary set of students and postdocs as diverse as molecular cell biology, physics, mathematics, engineering and chemistry. He inspired so many with his brilliance, enthusiasm, and passion for science and life. Generations of students, friends, and colleagues will never forget the fountains of ideas gushing out of him in the Brewed Awakening café on Euclid Avenue at the edge of the Berkeley campus, where George held informal group meetings almost every morning. George was also a generous mentor both professionally and personally to his students and postdocs in their advancement. He also served for more than two years on the Berkeley Division of the Academic Senate's Library Committee (2008-2011), helping colleagues from across campus to see how digital tools where changing the way the academy expanded knowledge. He retired from the University of California Berkeley in 2016.

For discovery of physical principles behind intracellular force generation in cell motility, morphogenesis and biological pattern formation.

Oster was very fortunate to be a part of the great generation of scientists who had audacity and freedom to move between disciplines. Early in his career, he worked on pathbreaking mathematical models of population dynamics and developed the first detailed theories of caste evolution among social insects (with Edward O Wilson) as well as theories of bifurcations to a hierarchy of periodic oscillations and chaos in simple ecological models (with Robert May). ... He was one of the first biophysicists who 'married' experiment and theory by developing computational models of a bewildering variety of molecular machines … [He] established a now common approach to solving biological puzzles by iterative cycles of modelling and experimentation. ... It would not be an exaggeration to say that Oster more than any other physicist or mathematician inspired our current understanding of molecular motors. ... [He] pioneered an important shift from conceptual to detailed and predictive mathematical models of biological systems.

[George was] a warm man who radiated the joy of doing science. He will be sorely missed by the scientific community and by his many friends who have been lucky to bask in the light of George's mind.

When I was coming up, things were a lot looser, jobs were easier. Now it's quite tight, and I think that people have to specialise in ways that I didn't have to. It's hard to get away with being a dabbler like me these days. I'm amazed at how lucky I've been. ... I think I just met the right people at the right time and worked on the right problems. I think of my career as a sequence of lucky accidents, of meeting people who were really smart. In fact, I think the story of my professional life is being able to work with people smarter than I am. ... When I was sailing in the engine room of ships, sweating, cursing my lot, and drinking coffee, not in a million years did I think I would be here. I remember Rod Laver, the tennis champion, once said to a reporter, 'I can't believe they pay me to do this.' Every day I look up at the Berkeley hills, and that's the way I feel. It is so much fun that I would pay them to do it. The idea that I would end up in Berkeley as a professor, let alone in the National Academy? It never crossed my mind.

... agree that his sharp intellect, physical intuition, and passion for scientific inquiry not only inspired us as scientists but also greatly influenced the way we conduct research.

Those of us who were lucky enough to work with George were forever shaped by his tenacious drive to understand the mechanisms underlying diverse biological phenomena. George often quoted Aharon Katchalsky, his postdoc advisor at Weizmann: "It is easier to make a theory of everything, than a theory of something." Modelling a biological system is indeed a daunting task and the "theories of something" that we have so proudly produced were sometimes invalidated by subsequent experiments. George, however, always insisted that despite the difficulties, we must focus on modelling specific systems because this modelling approach was more useful for advancing our knowledge of biology. Although validation by subsequent experiments was certainly rewarding, contradiction motivated us to revise and improve our models. With each iteration, we learned something new and got closer to the truth. While we worked with George, we became disciples of his approach to science. Now, as independent investigators, we try to continue his legacy by exposing our trainees to his style of research.

Quite a few mathematical fads have burst onto the popular press, promising great new insights into biology: information theory, catastrophe theory, chaos theory, and most recently complexity theory, to name just a few. All were quite useful in their founding fields (except complexity, which has proven slippery to even define), but all failed to live up to their hype, and faded from the public eye, slinking back to the specialties where from they arose, and where they still prove quite useful. But none has really contributed much to biology. As for chaos, it is a mathematical phenomenon that biologists were well advised to be aware of: that deterministic models could mimic random variation in some ways. But beyond that, I can't see that it has provided any startling biological insights. In physics and chemistry, mathematical phenomena have occasionally found amazing correspondences in the real world, but this has not happened very much in biology, where biological truths are not found in mathematics, but in measurements. Someday this may change, but I don't foresee it happening any time soon.