Prass et al. provide an analysis of the domain level diversity of bioaerosols from the Amazon rainforest. This dataset provides a window into an area of the atmosphere considered to be less affected by human pollution, which could help deconvolve what microbes are “naturally” in the atmosphere vs there by human-activities. It also provides a dataset that focuses on enact cells compared to sequencing methods that could also include eDNA. 

Major comments:

1) The manuscript would be improved by more discussion of other bioaerosol-microbe focused papers and how they relate to the current findings, such as Souza et al. 2019 (referenced but not discussed), Stern et al. 2021; Env. Sci.&Tech. Zhen, Sci. Total Environ. 2017; Yamaguchi, N. Sci. Rep. 2012. What is known about the different domains and their potential ecosystem roles or residence times?

2) The study seems to pause right when something unique to FISH (the future work proposed at the end of page 10) could be presented. It is unclear why only FISH was utilized. The current standard for airborne microbiology appears to be sequencing based (see papers in point 1), which provides higher taxonomic resolution than the domain level FISH analysis conducted in this manuscript. 16S rRNA qPCR is more robust than microscopy for counting as well. The manuscript lacks an explanation why sequencing techniques were not applicable for this system (why wasn’t FISH and sequencing done?) and should justify the decision to only look at domain level diversity (why weren’t more FISH probes used?).

What is the standard microbial density of a given aggregate? Are all domains found in physical association? Separate? How are they dispersed on the aggregate? This would be something a sequencing-only study could not provide and better justify the methods used. There are qualitative statements to this effect that could be expanded upon (pg 10 lns 13-16).

3) It is unclear why there is an improved understanding of bioaerosols from the number of bacteria present as opposed to total number of microbes present. For example, the paper states that these data provide constraints on mixing information, but it is not known if the bacteria are the same or different throughout. Is it the same population of bacteria that travel through the different heights? Or entirely different bacteria?

Minor and specific comments:

Methods - It is not clear how the determination for the genome size of all bacteria and archaea was made. Was this an average of all the genomes? How might this vary between cells?

The SD is very high compared to the average (Table 1). How many filters were counted? A supplemental table of each count conducted per time/height would be useful. It would be useful to know the variability in counting a given filter (variability in counts per field of view) separately from the deviation between filters counted for a given experimental filter (Table 1).

Man-made could be replaced with anthropogenic or human.

Figure 2 – why are there no “unknowns” for Mar 1 and Mar 2?

The manuscript by Prass and colleagues investigates the altitude distribution of bioaerosols in the Amazon rain forest. In summary, there are several strengths to this study: 1. bacteria, archaea, and eukarya were enumerated using fluorescent microscopy and FISH, 2. the site is an undisturbed forest environment, 3. collection was done at several different heights. There are also several weaknesses to this study: 1. only one week of data collection, 2. only one site of collection, 3. only one time/day of bioaerosol collection for FISH analysis, 4. Only one analytical technique (FISH) used to identify particles as bacteria, archaea, or eukarya.

While FISH is a laborious technique and it has been used before on some aerosols to distinguish bacterial species from each other, it has not been used for atmospheric bioaerosols and not to distinguish bacteria, archaea, or eukarya from each other. In particular, the ability to observe and enumerate assemblages of different composition (for example, assemblages that consist of only bacteria or assemblages that consist of eukarya and bacteria) is a clear strength of this technique. Figure 4H is a beautiful example.

The determination of absolute and relative numbers of bacteria, archaea, or eukarya at different heights in the Amazon rain forest is by itself another important result. However, it is clearly limited by the fact that collection was restricted to a one-time seven-day period and that the results were not compared with any other analytical technique. Because FISH was not used before to determine concentrations of bacteria, archaea, or eukarya at different heights at other sites, the results cannot be compared with other studies that used different techniques.

Of course, it is impossible to change the study itself at this point. It is the opinion of this reviewer that some improvements to the manuscript itself can make this an important and interesting contribution. The main recommendation is to more clearly acknowledge the limitations I listed above. In particular, the authors should not state at the same time that their data provide “unprecedented insights” and are “highly consistent with … previous studies”. The authors should instead acknowledge that the absence of an independent verification using other techniques, such as sequence-based techniques or qPCR, and the absence of similar studies performed at other sites during other season limits the ability to compare their results and verify the accuracy of their results.