deletions | additions       diff --git a/Discussion.tex b/Discussion.tex  index abb8f3d..6fa3db3 100644  --- a/Discussion.tex  +++ b/Discussion.tex 
...
\section{Discussion}   \subsection{Biomass Pool Size}  The goal of this study was to evaluate how changes in resource stoichiometry  affected the biomass pool size, membership and structure of planktonic and  biofilm communities. Our results suggest that C subsidies increased bacterial biomass in both plankton and biofilm communities as predicted. Carbon subsidies also resulted in decreased algal biomass in the plankton community, but there was no significant change in algal biomass of the biofilm communities among resource treatments. The changes in the biomass pool size that did occur were consistent with changing relationships (commensal to competitive) between the autotrophic and heterotrophic components of the plankton communities but not necessarily of the biofilm communities. \subsection{Biofilm and Plankton Alpha and Beta Diversity}   Beyond changes in the biomass pool size of each community we explored how  shifts in resource C:P affected a) the membership and structure of each  community, and b) the recruitment of plankton during biofilm community 
...
understanding the assembly of aquatic biofilms. First, biofilm community  richness was consistently higher than planktonic community richness  (Figure~\ref{fig:rarefaction}). Second, for the control, C:P = 10 and C:P = 100  resource treatments the membership and structure of the bacterial biofilm and plankton communities were more similar within a lifestyle (plankton vs. versus  biofilm) than within a resource treatment. However, for the bacteria in the highest C:P treatment (C:P = 500) both membership and structure of biofilm and planktonic communities at day 17 were more similar to each other than to communities from other treatments (Figure~\ref{fig:pcoa}). Third, C subsides acted differently on the algal and bacteria communities. Specifically while the highest level of C subsidies (C:P = 500) resulted in a merging of membership in the bacterioplankton and bacterial biofilm communities the same merging of membership was not observed for the algal biofilm and plankton communities which had distinct membership in all treatments.   We propose two potential mechanisms that could result in the increased  diversity of the biofilm communities relative to the planktonic communities.  First, it is possible that the planktonic community composition of our flow  through incubators was dynamic in time. In this case the biofilm community would represent a temporally integrated sample of the planktonic organisms moving through the reactor resulting in higher apparent alpha diversity (i.e. mass effects would be the dominant assembly mechanism). Second, the biofilm environment may disproportionately enrich for the least abundant members of the of the planktonic community. In this case it is probable that the biofilm would incorporate the most abundant members from the planktonic community (i.e. mass effects) but also select and enrich (i.e. species sorting) the lease abundant members of the planktonic community resulting in a higher level of detectable alpha diversity. The second mechanism would result if the biofilm environment represented a more diverse habitat including sharply delineated oxygen, nutrient and pH gradients that are not present in the planktonic environment. In this case the more diverse habitat would be able to support a more diverse community due to an abundance of additional environmental habitats (i.e. niches). We evaluated the first mechanism by comparing membership among the plankton samples taken 9 days apart (t=8 and t=17). While bacterioplankton communities were not identical between the time points (Figure~\ref{fig:pcoa}), communities within a treatment were more similar to each other between timepoints than any other bacterioplankton community (treatment or timepoint). In addition, the control and two lowest C treatments (C:P=10 and C:P=100) separated completely from biofilm communities in principle coordinate space (Bray-Curtis distance metric). This suggests that the biofilm community was not integrating variable bacterioplankton community membership, but rather selecting for a unique community that was composed of distinct populations when compared to the most abundant members of the plankton community. As noted above, in the highest C treatment (C:P = 500) the bacterial biofilm and plankton community membership had significant overlap at the final timepoint (Figure~\ref{fig:pcoa}). However, bacterioplankton membership for the highest C  treatment among timepoints (8 and 17 days) were also qualitatively as similar  to each other as any other community. Thus, variable planktonic community 
...
of the \textit{in situ} resource conditions were sufficient to alter the  relative abundance of the populations within each community. Second, an  analysis of the OTU relative abundance in biofilm and planktonic libraries  where OTUs are sorted by planktonic sample rank (Figure~\ref{fig:rank_abund}) shows that the least abundant members of the plankton community were routinely highly abundant within the biofilm community. This was true for both algal and  bacterial communities, at all treatment levels and both timepoints. While we  did not (could not) specifically measure niche diversity within the biofilm 
...
community composition and biofilm formation on glass beads placed for three  weeks in three boreal freshwater streams \citep{22237539}. While that study  system is markedly different than our study, the analyses and questions  addressed in each study were sufficiently similar to merit comparison. \citet{22237539} concluded that the biofilm community membership was most  likely driven by species sorting over mass effects. This is consistent with  what we report here. However, in the \citet{22237539} study the authors 
...
conducted in three boreal streams during snow melt when connectivity between  the terrestrial and aquatic habitats was high and potentially highly variable  depending on how hydrologic pathways differed among precipitation events. In  this study the source community was a marine intake located approximately 200m from the shore during July when communities are more stable over the 17 day period of the incubation. A separate study conducted in alpine and sub-alpine streams of the Rocky Mountains clearly showed that stream plankton communities reflected localized precipitation events and could be traced largely to sources of soil communities of drainages within the watershed \citep{22626459}. While planktonic communities in lake ecosystems can be linked to soil communities in the watershed, as residence time of the system slows the relative influence of species sorting increases. Thus, in headwater ecosystems stream plankton  communities can often be composed primarily of soil organisms \citep{22378536}.  In addition to the diverse source communities the \citet{22237539} study 
...
appears, however, that sample-wise bacterial biofilm rarefaction curves may  exceed the integrated planktonic curve upon extrapolation and most exceed the  integrated planktonic curve at sampling depths where data is present for the  biofilm and integrated planktonic library (Figure~\ref{fig:integ_rarefaction}). This result is consistent with our conclusion that temporal heterogeneity in the plankton was not sufficient to explain the higher diversity in the biofilm sample but would explain the relative differences between planktonic and biofilm diversity found in \citet{22237539} compared to this study. In addition, for this study, it is important to note that biofilm community  richness peaked at the intermediate treatment (C:P = 100) and appeared to  decrease over time although with only two time points it was unclear how pronounced this effect was (Figure~\ref{fig:rarefaction}). Since biomass of the plankton and the biofilm increased with increasing C subsidies the intermediate peak in OTU richness is consistent with a classic productivity-diversity relationship that has been shown for many ecosystems and communities both microbial and otherwise. However, as with other experiments with this result our experimental design did not allow us to tell whether resources drove productivity that drove changes in diversity or whether resources drove diversity which altered productivity. Rather we note that, as diversity decreased in the highest C treatment bacterioplankton and biofilm membership became increasingly similar. This suggests that environments that contain high amounts of labile C selected for fewer dominant taxa that came to dominate the biofilm community, overwhelming the species sorting mechanisms that appeared to dominate biofilm community assembly in all other treatments. Similarly, while  we did not measure extracellular polymeric substances (EPS), direct microscopy  showed that planktonic cells in the highest C treatment (C:P = 500) were  surrounded by what appeared to be EPS. Because biofilm EPS appeared also to  increase moving from the low to high C treatments (Figure~\ref{fig:microscope}) it is possible that more abundant planktonic cells were more readily incorporated into biofilms due both to increased "stickiness" of the planktonic cells as well as the biofilm itself. While we did not observe flocculating DOC which has been shown to dominate high DOC environments in nature, we did measure a substantial increase in DOC in the C:P = 500 treatment which was more than 2-fold higher than any of the other treatments. Thus additional adhesion of the plankton and the biofilm may also explain the merging of the planktonic and biofilm bacterial membership in the highest C treatment. \subsection{Lifestyle (biofilm or planktonic) Enriched OTUs}  There are only a few studies that attempt to compare biofilm community  composition and the overlying planktonic community \citep{Besemer_2007,  22237539, Jackson_2001, Lyautey_2005}. Those studies illustrate community composition among the two habitats are unique with very few taxa found in both.  This is consistent with our findings in this experimental system with a natural  marine planktonic source community. In addition, our study also evaluated algal community composition which showed a similar result suggesting that both the algal and bacterial biofilm communities form from phylogenetically unique organisms that exist in low abundance in surrounding habitat (i.e. the  plankton) but are readily enriched in the biofilm lifestyle. Most of the  biofilm enriched algal OTUs were \textit{Bacillariophyta} although there were 
...
the same taxonomic rank \citep{Schloss_2011}. Unfortunately, at higher  taxonomic resolution (e.g. Genus-level), groups did not possess a sufficient  number of OTUs to evaluate coherence between taxonomic annotation and  lifestyle. Similar to the richness results, we found the shape of rank abundance distributions between biofilm and planktonic libraries in our data (Figure~\ref{fig:rank_abund_shape} to be in contrast to that reported by \citet{22237539} although this may be an artifact of each study's different source communities (as discussed above). Carbon amendments did not affect algal library membership and structure to the same degree as it affected bacterial library composition. As expected, bacterial OTUs enriched in the high C amended mesocosm (C:P = 500) include OTUs in classic copiotroph families such as \textit{Vibrionacaea} and \textit{Pseudomonadaceae}. Interestingly, the one OTU depleted in the high C treatments is annotated as being in the HTCC2188 order of the \textit{Gammaproteobacteria}. HTCC stands for 'high throughput culture collection' and is a prefix for strains cultured under low nutrient conditions \citep{Cho_2004, Connon_2002}. \subsection{Conclusion}  In summary this study shows mechanistic links between large scale community  level dynamics and the underlying constituent populations that compose them. We  found that autotrophic pools and heterotrophic pools responded differently to  amendments of labile C as hypothesized. Notably while C amendments altered both pool size and membership of the bacterial communities we did not see similar dynamics within the algal communities. Planktonic algae decreased in response to C amendments presumably in response to increased competition from a larger bacterial community, however there was not a similar decrease in biofilm algal community. In addition membership of the algal communities between the plankton and biofilm lifestyles did not become more similar in the algae as it did for the bacteria in the highest C treatment. Consistent with a growing body of work our results suggest that complex environmental biofilms are a unique microbial community that form from taxa (both heterotrophs and autotrophs alike) that are found in low abundance in the neighboring communities. This membership was affected by resource amendments for heterotrophic but not autotrophic microbes and then only in the most extreme resource environment. This suggests that lifestyle is a major division among environmental microorganisms and although biofilm forming microbes must travel in planktonic form at some point - reproductive success and metabolic contributions to biogeochemical processes comes from those taxa primarily if not exclusively while they are part of a biofilm. Our results point to lifestyle (planktonic or biofilm) as an important trait that explains a portion of the exceptional diversity found in snapshots used to characterize environmental microbial communities in space and time.