Changes in land use and agricultural practices in tropical soils are t
hought to make major contributions to the global soil source of nitrou
s oxide (N2O) and sink for atmospheric methane (CH4). The variety of m
easurements of these gas fluxes in tropical systems are, however, limi
ted. From November 1992 until February 1995 we conducted weekly CH4 an
d N2O flux measurements within three grassland locations in western Pu
erto Rico that had not been cultivated for at least 25 years. Sites we
re located in either Vertisol, Ultisol or Oxisol soils that are typica
l of soils in the tropics. Plots within each site included non-fertili
zed controls and plots where fertilizer was applied at recommended rat
es (300 kg N ha(-1) y(-1) quarterly) for forage production in the regi
on. During the measurement period, CH4 uptake rates averaged 5.8 mu g
CH4-C m(-2) hr(-1) with no significant differences across sites. These
uptake rates were generally 10-fold lower than those reported for tro
pical forests. Fertilizer addition had a small negative affect on CH4
uptake in the Vertisol, tended to enhance CH4 uptake in the Ultisol an
d significantly decreased CH4 uptake in the Oxisol. Nitrous oxide flux
es averaged 16.9 mu g N m(-2) hr(-1) from unfertilized sites, with no
significant differences across sites. These background emission rates
were typically higher than those in temperate grasslands. In fertilize
d soils, N2O emissions averaged about 46 mu g N m(-2) hr(-1) in the Ul
tisol and Oxisol and 130 mu g N m(-2) hr(-1) from the Vertisol, about
0.8 and 3.3% of the N applied. Fluxes from fertilized sites were withi
n the range of emission rates observed for fertilized temperate grassl
ands. (C) 1997 Elsevier Science Ltd.