Ammonia (NH3) is the dominant gaseous base in the atmosphere and the principal neutralizing agent for atmospheric acids, yet remains one of the least well characterized atmospheric trace compounds. In particular, the spatial and temporal distribution of the background concentrations in terrestrial ecosystems and the importance of natural emissions from undisturbed soils and vegetation is poorly understood. This situation persists because of experimental difficulties associated with ammonia measurements, the rapid gas‐to‐particle conversion of ammonia in the atmosphere, and the capacity of native soils and vegetation to act as both source and sink for atmospheric ammonia. In the present paper, we attempt to summarize the current understanding of the natural sources and sinks for gaseous NH3 and the importance of natural emissions relative to anthropogenic emissions in the United States. We briefly review the physical and chemical processes that transform NH3 in the atmosphere, the major anthropogenic and potential natural sources of atmospheric NH3, and the techniques used to measure low concentrations and fluxes of atmospheric NH3. The available background concentrations and flux measurements of ammonia in natural ecosystems are then described and used to infer upper limits for the emissions of NH3 into the atmosphere from these systems. While the magnitude of both anthropogenic and natural emissions of NH3 remain uncertain, it appears that unperturbed terrestrial ecosystems are generally more important as sinks rather than sources for atmospheric NH3. However, net emissions are likely from many eastern forests and other ecosystems exposed to large inputs of atmospheric sulfate.

Nitrogen (N) fertilization of agricultural systems is thought to be a major source of the increase in atmospheric N2O; NO emissions from soils have also been shown to increase due to N fertilization. While N fertilizer use is increasing rapidly in the developing world and in the tropics, nearly all of our information on gas emissions is derived from studies of temperate zone agriculture. Using chambers, we measured fluxes of N2O and NO following urea fertilization in tropical sugar cane systems growing on several soil types in the Hawaiian Islands, United States. On the island of Maui, where urea is applied in irrigation lines and soils are mollisols and inceptisols, N2O fluxes were elevated for a week or less after fertilization; maximum average fluxes were typically less than 30 ng cm−2 h−1. NO fluxes were often an order of magnitude less than N2O. Together, N2O and NO represented from 0.03 to 0.5% of the applied N. In fields on the island of Hawaii, where urea is broadcast on the surface and soils are andisols, N2O fluxes were similar in magnitude to Maui but remained elevated for much longer periods after fertilization. NO emissions were 2–5 times higher than N2O through most of the sampling periods. Together the gas losses represented approximately 1.1–2.5% of the applied N. Laboratory studies indicate that denitrification is a critical source of N2O in Maui, but that nitrification is more important in Hawaii. Experimental studies suggest that differences in the pattern of N2O/NO and the processes producing them are a result of both carbon availability and placement of fertilizer and that the more information‐intensive fertilizer management practice results in lower emissions.

Legumes may be used to reduce fertilizer inputs, provide ground cover for soil erosion protection, and reduce residual soil nitrates. Little data are available to guide a producer in selecting a legume that matches water and temperature conditions and produces acceptable growth and N 2 -flxation rates. The purpose of the experiment reported here was to provide such information. After a 14-d germination period at 20 °C, we grew eight legume species at soil temperatures of 10,20, and 30 °C (± 2 °) for 105 d in a greenhouse. Daily water use was recorded, and plant dry weights were measured every 21 d. Initially annual species such as soybean [Glycine max (L.) Men.] and fababean (Viciafaba L.) grew fastest at all temperatures. For the first 42 d at 10 °C, white clover (Trifolium repens L.), crimson clover (Trifolium incarnatum L.), and hairy vetch (Vicia villosa L.) also exhibited rapid growth. With warmer temperatures and longer growth periods, lespedeza (Lespedeza stipula L.) as well as soybean grew rapidly. At 20 and 30 °C, soybean growth was often more than double that of most other species. Growth was maximal at 10 °C for fieldpea (Pisum arvense L.), hairy vetch, and crimson clover; at 20 °C for fababean and white clover; and at 30 °C for soybean, sweetclover (Melilotus alba L.), and lespedeza. Average water-use rates and water-use efficiency generally paralleled growth, except water use increased and water-use efficiency decreased as soil temperature increased. These results indicate that growth and water use differed greatly among legume species, that each legume species has its own characteristic optimum growth temperature, and that this optimum temperature for a given species may change as growth progresses. P RODUCTION COSTS plus soil erosion and water quality concerns have resulted in many farmers and agronomists evaluating the potential benefits of using cover crops (Power and Follett, 1987). A legume provides ground cover to reduce erosion potential, reduce fertilizer N inputs and costs, and deplete residual soil nitrates and water, thereby reducing potential for ground water contamination. Use of cover crops may

Wben selecting a legume cover crop, one should know relative N· fixing and N uptake capabilities, as well as growth and water use characteristics, to identify the species best adapted to the growth pe· riod and soil temperatures (season) during which the cover crop is grown. We provide information on these characteristics for eight in· oculated legume species at soil temperatures of 10, 20, and 30 °C. Plants were grown in constant-temperature water baths in a green· house for 105 d after establishment in 1.1 kg of Alliance silt loam (fine silty, mixed, mesic, Aridic Argiustoll) per pot. Plant samples were taken every 21 d for determinations of dry weight, total N uptake, and N, fixed (isotope dilution method). Water use was measured daily by weighing. Total N uptake and N, fixation were usually greatest for large-seeded annual species during the first 42 to 63 d of the experi· ment. At 10 oc total N uptake and N, fixation were greatest for hairy vetch (HV), VICia yil/osa Roth and faba bean (FB), VICia faba L. At later sampling dates, N uptake and fixation for white clover (WC), Trifolium repens L., was also relatively high. At 20 oc, soybean (SB), (Glycine max (L.) Merr.) exhibited outstanding growth and N uptake throughout the 105 d. For the first 42 d, FB performance also was superior to other species. At 30 oc, N uptake and fixation by SB was more than double that of any other species at all sampling dates.Quantity of N, fixed per unit water used was greatest at 10 oc for WC, followed closely by HV and field pea (FP) Pisum saJiyum L.; at 20 oc, SB followed by WC and lespedeza (LD), Lespedeztl stipulacea Maxim.; and at 30 oc, LD followed by SB. Our results suggest that under many situations (early spring) some grain legumes, such as SB and FB, may be a better cover crop than many species commonly used.