The consensus is a skimmer is needed, but that notion was first stated as a “must” when carbon dosing for aquariums was first described. You should be aware that the edict was based on little experience and no measurements. It seems like a reasonable recommendation though.
I have run 2 liter carbon dosing experiments and have seen little oxygen consumption. I cannot speak to CO2 production except to say that the pH is stable. I have more experiments to conduct. I can only wonder at this point whether the skimmer requirement is correct.
What I think everyone is forgetting in this debate is that you will be increasing the carbon dose (vinegar, ethanol, glucose, whatever) gradually until an effect is observed (hence the requirement to check nitrates often). Oxygen consumption, carbon dioxide production, bacterial biomass and bacterial exudates by bacteria will increase slowly. Carbon dosing does not produce bacterial blooms until you get to high doses, I think somewhere above 0.5 mL/gallon. If you are going to dose without a skimmer, be very observant of the condition of your aquarium. If you don’t have math phobia, you can calculate the maximum amount of CO2 produced and oxygen consumed per vinegar dose and prove to yourself the size of the worse case scenario you face for each dose increase.
As for skimmer removal of small creatures, it is a theoretical possibility. I have put skimmate under the microscope, and yes, living protozoa have been harvested. I have occassioning seen pods in the skimmate too, but, it is not teeming with them. I would say that to date, the skimmer has not been identified as a killer of aquarium microscopic wildlife.
