The purpose of this book is to expound a long chain of hypotheses: (a) thatHomo sapiens evolved from a previously unremarkable hominid as the result of two crucial mutations, the first, between 0.6 and 0.15 million years ago, affecting phospholipid metabolism and resulting in a sudden rise in brain size and an enormous increase in synaptic complexity, the second, perhaps 150 000 to 130 000 years ago, involving the phospholipase A₂ cycle and producing, as a package deal, both the technological and artistic creativity and the ruthlessness that are the essence of our humanity, and also schizophrenia, bipolar illness and dyslexia; (b) that thereafter the balance between the beneficial and harmful elements in this package depended on the essential fatty acid (EFA) content of Homo sapiens' diet: so long as this was high, psychotic illnesses were mild and inconspicuous, but with the advent first of agriculture and later of urbanisation, psychoses became more common and more florid; and (c) that re-establishing an adequate intake of EFAs is the key to the prevention and treatment of these disorders and that eicosapentaenoic acid is probably the crucial substance. Not for nothing was Horrobin the founding editor of the journal Medical Hypothesis.

Perhaps because the author has already lost hope of influencing the scientific establishment, the book is written for a general readership. Much of its text consists of descriptions of basic clinical, genetic and biochemical processes, but it is written in a fluent, engaging style. I do not know enough about anthropology, lipid metabolism or human genetics to know how plausible his various hypotheses are from the vantage points of those disciplines, but I do know that Horrobin's key assumption that there is a striking excess of highly intelligent, creative high achievers in the families of people suffering from schizophrenia or bipolar illnesses is far from proven. The idea goes back at least to Galton, but apart from Karlsson's studies in Iceland, it is based almost entirely on clinical impressions, not on defined populations and certainly not on blind ratings.

Reactions to this book are likely to be very diverse. It will probably be acclaimed with delight by many patients and their families, because it gives them hope and restores their dignity, and dismissed as fantasy by many psychiatrists and neuro-scientists, because it has almost no points of contact with contemporary aetiological theories and research. As Horrobin disarmingly admits, it may all be a Kiplingesque ‘Just-so story’ but, as he also points out, it does contain testable elements.

In my view, the clinically relevant elements in his chain of hypotheses ought to be taken seriously, if only because our understanding of the causes of schizophrenia and bipolar illness, and our ability to help people with these disorders, have hardly advanced in the past 40 years. It may seem unlikely that schizophrenia is fundamentally a disorder of phospholipid metabolism exacerbated by a dietary deficiency of EFAs, but the example of the Mensa & Dougie mice demonstrates that modifying a single gene in theN-methyl-D-aspartate phospholipase A₂ pathway can produce a massive increase in intellectual performance, and if eicosapentaenoic acid or some other EFA is indeed an effective therapeutic agent in the treatment of schizophrenia that would be very, very important.