Upon careful reading, I come to the realization that I am not qualified to review this manuscript, due to its heavy reliance on mathematics. Comments below are suggestions only.

Many technical jargons were used without being clearly defined first. For examples, effective population size, selective value, founder effect.

Ln57-59: this is exactly what happens as infections spread from one cell to the next. We now know that if a very small number of viral genome copies (say under 10) invade a cell, the infection is unlikely to be successful. Thus, for most viruses productive cellular infection requires cooperation among large numbers of viral genome copies co-entering the same cell. Nevertheless, regardless of the number of genome copies a cell internalizes, only very few of them have the chance to replicate to yield progeny genome copies (intracellular bottlenecking). The same intracellular bottlenecking process repeats itself when these progeny genome copies enter the next cell, so that again only very few of them yield progeny. This has been well documented – e.g. Grubaugh et al Cell Host-Microbe 2016.

The authors should consider discussing Muller’s ratchet in the context of new findings by others over the last 10 years. It is true Muller’s ratchet has the potential to cause some isolated subpopulations (cell lineages in an infected individual) to accumulate deleterious mutations and potentially go extinct, but the same process could also allow other subpopulations (intra-host cell lineages) to become more fit. In the case of viruses, the subpopulations that succumbed to Muller’s ratchet exist for a time too brief for us to study. This is simply what natural selection is about.   

Ln60-63: As discussed above, I cannot agree that one virus particle is enough to initiate a successful infection in an intact plant. I challenge the authors to demonstrate that. Earlier studies came to such conclusion by measuring the progeny genome copies that replicated to detectable levels. However, just because very few viruses end up producing detectable progeny does not mean very few entered a cell/tissue/plant.

Ln207, define “100% infectivity”.

Ln142-145, unclear. Were the up to 10 lesions mixed before grounding? How did you ensure all of the 10-microliter-extract, or all virions in it, invaded tobacco cells? The issue here may indirectly prove that you need a high number of virions to enter each cell for successful infection to occur (the content of a single lesion may become so diluted or wasted in the preparation and inoculation processes so that not enough viral genomes can invade a single cell together.   

Ln207-208, Did all lesions invariably lead to poor infectivity? Do you sense something wrong here?

Ln210, 212, unclear. “serial passages” – all in quinoa?  

The introduction is clear about the theory and objectives behind the experiment. It aims to investigate the survival of viruses infecting plants despite the action of Muller's ratchet and transmission bottlenecks. The references cited are sufficient. I also highlight the study's main contribution in that it regards a whole organism (in this case, the quinoa and tobacco plants) rather than culture cells, traditionally used to explore the same questions. The research questions are straightforward, though hypotheses or predictions still need to be included. In general, there isn't a clear answer to the title question; perhaps it could be more explicit that they are conducting an experimental approach in the title.
The methods are clear enough to allow for experiment replication, though it took me a while to understand them. Figure 1 could be further improved to make understanding the passages and differences between the Ne expected sizes easier. For instance, it could be helpful to include the approximate—number of generations. In general, quinoa plants are infected to create lineages, a reduced number of individuals, and allow for the accumulation of mutations. Then viruses are transmitted to a natural host of tobacco plants that will show a systemic infection. This approach is consistent with the question. Further, the fitness measure would be the infectivity measure in the tobacco plants, and this specifically would make many assumptions. I would expect that the plants measured are all clones and not influenced by external environmental factors during the 9dpi, which I doubt was the case.  
Further, the authors tested their experimental results by comparing them with a model. The model is explained in terms of the parameter values, and the authors discuss them quite well. However, it could also be helpful for the reader to have a background in the choices of these ranges, i.e., why the replication (r)= [1,50] and why they chose 200 replicates. The size of the bottleneck or the initial infection number of individuals is better covered but less for the rest. Additionally, the readers might benefit from having more information about the virus biology (what is in the literature), i.e., the replication time (generations), the mutation rates, genome recombination, and epistasis. The model scripts run and produce consistent values with the reported in the paper. I am not surprised the author disagrees with their model and the experiment. The model cannot reflect all the complexity involved, as they recognized in the discussion ( for example, lines 292, 359). The model could recognize the limitations more explicitly and still show how valuable it is to try to model the transmission as it is done. Perhaps because the model is missing the reality of the number of infective particles in every passage, I understand the difficulties of measuring this before a passage, if it is possible in-vitro in this virus. The discussion also needs some specifics. For instance, when they mention the genetic properties of viral particles in lines 274-275, what exactly do they refer to? Haplotype load? Population load? Other?
And finally, in line 359, the authors mention cooperation. How could this happen? Do they mean epistasis? or host-pathogen interactions? What about the replication system of the virus? Are they exempt from progeny skew? Especially if the founder population is small. I also would be careful to claim mutation rate changes (line 411) as these are not measured here. A good question remaining is the role of mutation rate and mutation rate changes in these experiments. If it happened, what is the cause? how do you measure it experimentally?