s contribution confirms what many had suspected: that the rate of CO 2 emission to the atmosphere from tropical deforestation is substantially larger than what it was in 1980, the year on which previous analyses of the role of tropical deforestation in the global carbon cycle have been based. Houghton estimates a likely 1989 emission of 1.5-3.0 x 1012 kg C, compared to a 1980 emission of 1.0-2.0 x 1012 kg C using the same methodology and assumptions. This increase is a direct consequence of a dramatic increase in rates of deforestation for a variety of social, political, and economic reasons. The most serious consequence of this deforestation in my opinion is not its effect on climate or atmospheric carbon dioxide, but the massive species extinctions -a biological holocaust -which it implies. Absorption of atmospheric CO 2 by the oceans will remove about 85% of the emitted CO 2 within a few hundred years, dissolution of marine carbonate sediments will remove another 10% or so within a few thousand years, and silicate weathering will take care of the rest within about 100 000 years, which is a very short period of time from an evolutionary perspective. Species extinction, in contrast, is irreversible.
This having been said, there are a number of important implications of Houghton's contribution with regard to the immediate issues of atmospheric pollution, understanding the carbon cycle at decadal to century time scales, and policy responses. The first of these is addressed by Keller et al. (this issue): an increase in CO2 emissions due to burning of forests implies an increase in emissions of CO, NOx, and other trace gases. Andrea et al. (1988) estimated that 5-20% of the carbon in combusted material in burning tropical forests is released as CO rather than as CO2, and about half of the carbon released in 1989 by deforestation appears to have been a result of burning in that year. The emitted CO is ultimately oxidized to CO2 (thanks to the presence of OH in the atmosphere), but not before a number of far-reaching chemical effects, including production of ozone (a greenhouse gas) and indirect extension of the atmospheric lifetime of CH 4 and the CFCs (further greenhouse gases).
One of the persistent problems facing Earth System scientists is the elementary task of balancing the carbon cycle, that is, accounting for the difference between anthropogenic emissions and the observed rate of increase of atmospheric CO 2 in terms of plausible sinks of CO 2. The rate of atmospheric CO 2 increase during the period January 1986-January 1990 was about 3.5 x 1012 kg C yr -1. Fossil-fuel emissions averaged about 5.5 x 1012 kg C yr -1, which, combined with Houghton's likely deforestation emission of 1.5-3.0 x 1012 kg C yr -1, gives a total anthropogenic emission of 7.0-8.5 x 1012 kg C yr -1 and requires sinks of 3.5-5.0 x 1012 kg C yr -1. A recent study by Tans et al. (1990) indicates that the oceans were taking up