1 last question: if I do a 5g batch, then harvest the yeast and immediately do another identical batch, is there a general rule I could follow as to how much of the yeast I should reuse? I know it won't be as lively as the 1st batch, but others are saying that the cell count increases by 5x or 6x, so I assume I'd be using just a portion of it? Half, maybe? A quarter? (That's just a SWAG.)
There are a variety of ways people reuse the yeast at the bottom of the fermenter. The first, easiest, and laziest is to just pour your next batch directly on top of the trub left at the bottom of the fermenter. While this would most likely be a significant over pitch, you are also carrying over an extra ripe cake of trub which is not yeast or composed of dead yeast cells. It works for a repitch or maybe two, but sooner or later you will start to build up other off-flavors from the decaying trub and dead yeast cell autolysis, not to mention that other stray bacteria colonies would have a field day on the left over cake.
The second is to leave behind a fair amount of beer and use that to slurry up the yeast cake. Allow it some time (30 minutes to an hour depending upon the flocculation quality of the yeast) to settle and pour off the suspended yeast/beer slurry into sterilized or sanitized pint or quart jars. You can then refrigerate the jars and collect the yeast cake on the bottom. This will be a combination of yeast and light trub and should be white to cream in color. When you have a settle yeast cake you can estimate the concentration of yeast and density of the cake by assuming that the compacted yeast has a count of approximately 4.5 billion cells per ml. Depending upon how much trub you have mixed in, you can estimate a percentage of yeast to trub and discount this figure by your estimate to get close to a reasonable amount per ml of cells. Multiply this out by what you need to pitch and you have the volume you should measure out for your next batch. The beer in the solution should prevent other contaminants from growing. I would guess from my notes that this amounts to approximately half of what I collect in a following batch, though much depends upon my patience and available time which determines the cleanliness of the collected yeast cake.
The third is similar to the second except that instead of leaving a fair amount of beer behind, add a quart of sterilized water to slurry up the yeast/trub into. Proceed as above. This is my usual means of re-use of the yeast. I find that the resultant yeast cake is cleaner and brighter and healthier. The fluid is always diluted beer, so there is some alcohol present as a mild preservative though not enough to prevent issues if you carry over any contaminants from the previous batches or collected during the transfers.
Another method that I also use is to over build the starters. Since I do cell counts, I can remove from the starter the amount that I need for a batch and then either store or propagate more from the remaining yeast which is left over. This is becoming more popular as you don't have to worry so much about any character from the previous batch carrying over to the next batch. People feel that there is less chance of contamination, though I have not seen any difference in my efforts.
Overall, I use a combination of the last two methods described. I find that I get better results from harvesting most of the ale strains and then building up with starters once I get to the third or fourth generation (repitch) as that generation is where most of the strains I use really hits its stride. I've gone as many as 9 generations, but have found too many off-quality cells and less uniformity in cell appearance. I will also do a culture plate of strains that I intend to use again later in the same brewing season. This has helped me keep strains that I start with in the fall to reuse in the spring surrounding my run of lagers. It also starts me off with clean colonies of yeasts without worrying too much about contaminants. In the 4 years I have been doing this, I only came across one strain that showed any significant contamination and needed to be tossed, so I count myself lucky in that regard.
Lager strains (including hybrids such as WY2112) are a bit different. Here constant harvesting and repitching leads to issues with flocculation and settling. So I start with splitting small batches between two strains (usually a Helles or festbeir type of recipe) to grow up a base of yeast. I can repitch or propagate from those initial collections to fill in the next 10 to 12 brews by alternating the use of the two strains. Since I generally want my lagers to be clean and crisp, working from this initial harvest and using starters from that point to propagate further seems to keep the risk of mutations and off-flavors showing up.
Such reuse requires planning. I will use 8 different strains this year for approximately 34 to 36 different brews. Between the cost of DME, yeast nutrient, and initial cost of the yeast packs, I estimate that my yeast cost runs to just over $3/brew versus $7 a pack. This does include a few one-offs (such as WY1007 for an Altbier and WY2298 for a Czech Pilsner) so that tilts the picture a bit on the number of strains versus reuse.