A first step in developing understanding of relationships between gaseous precursors and sulfate particle mass and number concentrations is to examine the natural S cycle in regions minimally influenced by anthropogenic sources. The cycle of DMS is of particular interest, because of its potential link to indirect climate effects of aerosols; a number of studies have been aimed at quantifying DMS sea-to-air fluxes and yields of gaseous and particulate S products. The molar ratio in particulate matter of methansulfonate (from methanesulfonic acid, MSA, a DMS oxidation product) to non-sea-salt sulfate (MSA/NSS) has been investigated as a possible tracer for DMS-derived sulfate. For example, aerosol nitrogen and sulfur species were measured in Antarctica by Savoie et al. [1992]. MSA, ammonia and NSS exhibited similar seasonal cycles, with summer maxima; regression analyses indicated that the dominant source of NSS was DMS, with aerosol MSA/NSS ratios of about 0.31. Although early studies of Pacific aerosol from mid-to-low latitudes had suggested this ratio was near 0.065, subsequent work found that the ratio increases with latitude, possibly through the temperature dependence of the DMS oxidation chemistry ( Bates et al. [1992a] and Li et al. [1993]).
New particle formation is thought to occur via gas-phase reaction of
SO
with OH in the presence of water vapor to form H
SO
, a
species with very low vapor pressure that partitions to the aerosol
phase. Although laboratory studies suggested that SO
is the major
DMS oxidation product, the observations of Bandy et al. [1992]
were consistent with low yields of SO
and suggested that other
pathways were important. Lin and Chameides [1993] suggested an
alternative oxidation scheme that led to a stronger source of sulfuric
acid and supported the idea that homogeneous nucleation in the marine
boundary layer could serve as a source of CCN.
Methanesulfonic acid could contribute to particle formation. However,
modeling ( Kreidenweis et al. [1991]) suggested that MSA
preferentially condenses on existing particles and does not nucleate,
and that new particle formation from sulfuric acid occurs if
preexisting surface area is not too large. These ideas received
support from several observations. Pszenny [1992] measured
particle size distributions of MSA in the tropical marine boundary
layer. MSA was distributed according to effective surface area in
particles larger than 0.5 
m, consistent with condensational
growth, while smaller particles were primarily sulfate. From
observations of aerosol size distributions, Clarke [1992]
concluded that DMS oxidation products most likely grow existing
aerosol. However, Covert et al. [1992] report observations
detecting a burst of new particle formation in the marine boundary
layer in a case where particle surface area had become low.
It has been postulated that an additional sink for SO
in the
marine boundary layer is uptake and oxidation in sea salt
aerosols. This mechanism could be important to submicron aerosol
nucleation and growth processes because it would remove sulfate
precursor gases from future interaction with that population (
Luria and Sievering [1991]). Field studies ( Quinn et al.
[1993] and Sievering et al. [1991]) showed that a major fraction
of the non-sea-salt sulfate mass was associated with large (>0.9
m) sea-salt aerosol, supporting the potential role of sea-salt
aerosols in heterogeneous S oxidation in the marine boundary layer
(MBL) ( Chameides and Stelson [1992] and Sievering et al.
[1992]). Further evidence was obtained in the single-particle
observations of McInnes et al. [1994], who detected sulfur
enrichment and chloride depletion in both submicron and supermicron
aerosols from the remote marine Pacific. Their study also suggested
that sea salt playes a larger role than previously thought in the
accumulation-mode number concentration, indicating that sea salt may
act as a ``seed'' for the uptake and oxidation of S gases and vapors.
Clarke [1992] measured aerosol in the remote free troposphere
(8--12 km) between 70 
N and 58 
S. Low aerosol mass
concentrations were correlated with high number concentrations of
ultrafine, volatile aerosol, consistent with a sulfuric acid
composition. These observations suggest that clean regions of the
middle and upper troposphere act as significant source regions for new
atmospheric nuclei. Clarke [1993] postulated that, as reactive
gases are pumped to higher levels of the atmosphere by convective
clouds, the atmosphere is simultaneously cleansed of preexisting
aerosol, creating optimal conditions for this type of new particle
production. The nuclei have a long enough lifetime to allow them to be
mixed throughout the troposphere (including into the boundary layer),
where they may serve as the source of aerosol and CCN. Such a
mechanism would compete with new particle production in the MBL from
oxidation of S species and could explain why relationships between DMS
emissions and CCN concentrations have been difficult to
establish. Clarke [1992] also suggests that increased production
of dust and carbonaceous aerosols by anthropogenic activity could
increase aerosol surface area and suppress nucleation and lifetimes of
ultrafine aerosol in the free troposphere, with consequent
perturbations to the source strength of total aerosol and CCN.