Is this video of any use?
http://www.bobvila.com/BVTV/HomeAgain/Video-0512-03-0.html

Is this the basic concept I should be shooting for? How does drainage work after the water goes through the rocks and meets the polystyrene?

This question comes up regularily on the forum. Here is a summary:
Greenhouse Floor Insulation

From Alberta, Canada Dept. of Agriculture:
"The floor is covered with white plastic film to seal-off the soil from the greenhouse environment, reducing the problems associated with soil-borne plant diseases and weed problems." (Note: greenhouse floor insulation is not recommended in Canada)
http://www1.agric.gov.ab.ca/$department/deptdocs.nsf/all/opp2892#5

From Rutgers University:
"Two feet deep perimeter insulation (two inches thick) is installed to reduce the heat loss through the greenhouse floor." (Note that this insulation refers to the below ground area directly under the side walls only and not insulating the floor area)
http://aesop.rutgers.edu/~horteng/openroof1.htm

From several published reports:
Although the amount of heat lost in a greenhouse may vary a little in different environments, the published data is surprisingly similar. Averaging the research of USA and Canada publications, the typical heat loss is: roof-65%, walls-18%, infiltration (leaks) - 15%, perimeter walls- 2%, and floor (insignificant heat loss). One example is:
http://ohric.ucdavis.edu/Newsltr/fn_report/FNReportSp82.pdf

Greenhouse heating calculators include the following factors- roof area, walls, and ends. Floors are not a significant heat loss area.
The commonly used hobby greenhouse calculator even specifically mentions to exclude the floor in the calculations.
http://www.littlegreenhouse.com/heat-calc.shtml

Finally, warm air rises. Taking care of the factors that do matter is much more cost effective than spending money insulating the floor of the greenhouse.

Thank you for the information!

I've been doing tons of reading today, and it seems like insulating just the sides of the greenhouse floor around the perimeter of the foundation with 2" polystyrene about 2' deep is the way to go.

Now for the actual floor. I think that trying to be able to plant directly in the floor is just not feasible (of course, I could be wrong!). Maybe I should just put a weed barrier down about 1', put about 4" of sand over that, then 4" of pea gravel over that. Does this sound like a smart choice? If not, any other ideas?

Glad the information was useful. As for the greenhouse floor, I agree that trying to grow directly in the soil is difficult. There can be problems with irrigation, weeds, diseases and nematodes. For your weed barrier-sand-pea gravel idea, I have a suggestion for something easier and probably much less expensive. I use simply 6 inches of 3/4-1 inch coarse gravel directly over the native soil. The drainage is very good, no weeds seem to grow in the coarse material, and it is better to walk on than pea gravel. I do have an area outside with pea gravel, and the weeds are flourishing. My house is also full of the pea gravel that sticks to my shoes when I come inside. I am not sure what the purpose of the sand layer would be so I do not have any comments on that.

Cuestaroble, thank you for posting that link again.

From a post I made last month:
The perimeter heat loss is what you prevent by insulating vertically around the foundation. I've often wondered if that study showing 2-3% perimeter loss was done in a warmer climate. There is a way to calculate perimeter loss. This paper and this paper indicate that an uninsulated perimeter loses 0.8 BTU/linear ft-degree F-hr whereas an insulated perimeter loses cuts that in half to 0.4 BTU/linear ft-degree F-hr. I haven't taken a really close look at the numbers, but some quick calculations show that for a smallish, 6x8 GH, at 55F, avg outside/ground temp of 30F, the heat lost out the perimeter amounts to about 13% of the total heat lost. That's way more than the 2-3% in the studies discussed here previously.

The link you cite is for California greenhouses. In Missouri and Colorado, would you agree that the figure could be higher?

The base under my under my GH is as well insulated as the perimeter and while my desire and need for a warm GH floor may not be shared by others, the fact that my GH floor stayed barefoot warm all winter didn't hurt, because as has been noted, over and over again: warm air rises.

Also, my GH heater is mounted very low, so the hot air blows directly onto the glazed quarry tile floor, warming it, while the insulation under it holds the warmth in, so I still believe the under-floor insulation contributed to my steady 72 throughout the bitter cold, with very low heating bills.

The type 150 or heavier 250 that would be used under and/or around a GH is not permeable, so requires a drain, but if was sloped to an open space in the center, water would find it's way through gravel and down to that channel.

However, such an arrangement would preclude planting directly into the ground, as opposed to using large floor pots.

stressbaby,
Your calculations for the 6x8 are very interesting and accurate using the parameters given. I believe there may be several factors involved in the difference between 2-3% and 13%. A colder soil does give a higher perimeter loss as you mention. But in many places, the ground,especially inside the greenhouse, would probably not be as cold as 30 deg. Also, the published 2-3% is calculated for larger, commercial sized structures. In recalculating using a larger GH and a 45 deg. ground temp, the perimeter loss comes to the 2-3% figure.
Finally, I guess this supports the general idea of using perimeter wall insulation for some heat savings, but insulating the entire floor area is a different matter.

"The link you cite is for California greenhouses. In Missouri and Colorado, would you agree that the figure could be higher?"(stressbaby).
Actually, the link was just one example that happens to be on the internet. The figures I used were an average of several publications. One of my sources:"Energy Conservation in Ontario Greenhouses" give the following for Canada: heat loss- roof 61%, sides 27%, infiltration 10% and perimeter wall loss 2%.(no mention of floor).

Interesting...a study from Canada, and still only 2% perimeter wall loss...I need to play with numbers for larger greenhouses...thanks.

Regarding the floor...I went through probably half-dozen sites with formulas for calculating GH heat loss...they were the source of the BTU/linear ft-degree F-hr perimeter heat loss figures I quoted above. In none of those links was there any mention of heat loss through the floor.

I'll bet that as the size of the GH goes up, the percentage of the heat lost through the floor drops. The real question then becomes, at what point does it no longer become cost effective to insulate the perimeter...hmmm...