I must retract certain statements near the beginning and end of Note 295 on sea level rise, page 69b, particularly (a) the statement `that surface temperature changes penetrate into the ocean at a rate governed mainly by the rate of... diapycnal mixing, or mixing across the ocean's stable density stratification', and (b) the suggestion that the timescales are `indifferent' to circulation patterns. I must also retract the statements in Note 258, page 65b, that, if excess greenhouse forcing were constant, sea levels would rise `at fairly near a constant rate' (line 2) or at a rate `constant, to a first approximation' (line 16), for a century or more. All these statements pretend to a more precise knowledge than is available today. In particular, the diapycnal mixing discussed in Note 295, while part of the story, need not dominate, especially in the first few centuries of sea level change. It might even be relatively unimportant -- especially in the first few centuries -- if the geographical patterns of surface warming and upper ocean circulation were to develop in certain ways, so as to favour the transport of heat anomalies along, rather than across, the sloping stratification surfaces. In today's oceans, water can reach depths of order 2 km or more by moving sufficiently far along such surfaces, starting near the top of the ocean at certain high latitude locations; and this is another process having a multi-century timescale, a point that to my great embarrassment I overlooked. See, e.g., J. A. Church et al., `A model of sea level rise caused by ocean thermal expansion', J. Clim., 1991, 4, 438-456. The upshot is that my main point about sea level rise and its timescale, as made on page 47a and in Note 258, is almost certainly correct for practical purposes except that the precise rate of sea level rise will depend not only on the strength of excess greenhouse forcing but also on hard-to-predict changes in surface temperature patterns and upper-ocean circulation patterns, adding yet more to the uncertainties. On top of all this, some researchers consider that the West Antarctic Ice Sheet, much of which is grounded below sea level, might melt faster than predicted by current models of ice flow dynamics, possibly adding something like an extra metre per century to the total rate of sea level rise. West Antarctic ice volume above sea level corresponds to about 5 metres of sea level rise. The fog we are driving in is still denser than I thought. I am grateful to Trevor McDougall, Sarah Raper, and Stephen Schneider for making me aware of the mistakes in my reasoning. (Thus I have inadvertently illustrated my own point, made on p. 201 of Part I, about how, by the omission of trivialities, `one can fail to detect actual mistakes in one's own thinking'.)

I take the opportunity to call attention to Notes 132 and 177 of Part III, which correct slips in Parts I and II. At the end of Note 132, `bacterial populations' should of course read `bacterial and viral populations'. At the end of Note 142, `a year or so' should read `some months'. Further to Note 245, see the detailed analysis by a respected economist and economic historian, P. A. David FBA: `From market magic to calypso science policy: A review of Terence Kealey's The Economic Laws of Scientific Research', Research Policy, 1997, 26, 229--255. I thank John Ziman for pointing out this reference, which politely, firmly, and meticulously exposes the `flawed economic logic' in Kealey's book, `the telling of stories that blatantly disregard the evidence'.

I also made a complete fool of myself regarding plasmids. See Note 132 of Part III.