So I have been searching through the archives on the details of Garcinia pollination, and the information in this thread and others had not really clicked with me. For example, the notion that you could have a dioecious species - by the very definition, male and female flowers on separate plants - and not have to worry about having multiple plants. After all, even if the females would be facultative agamosperms and produced apomictic seeds, what if you got a male plant? A male plant won't produce any fruit.
After digging through some papers this evening, I think I understand what's going on - but correct me if I'm wrong. The part that clued me in was in a paper on G. hombroniana:
Species with sporophytic agamospermy such as Garcinia commonly also undertake regular sexual reproduction (Richards, 1986, 1990). For these, it is to be expected that a fertile diploid status will prevail, as it does in G. hombroniana. The high interspecific frequency, and the lower intraspecific density of plants in tropical habitats may select for agamospermous reproduction as a consequence of poor pollination, particularly when, as in Garcinia, dioecy has also arisen, perhaps in response to heavy female reproductive loads. The present work suggests that pollination is inefficient in G. hombroniana even under optimal conditions, and more than half the seeds set will carry asexual proembryos. Relatively few males will result, and this will exacerbate problems of pollen travel.
In short...
* There are separate male and female plants in most (but not all) Garcinia species (but there's a lot of variants to this pattern)
* The males can't fruit, only produce pollen. The females can reproduce sexually or asexually (cloning)
* Since the clones are exact copies of the parent (female) plant, they're also female. But if there's male pollen in the area, then flowers can be sexually fertilized, and roughly half of those seeds will be males.
* Since sexual reproduction is preferred, and designed to be efficient, so long as there's at least *some* males in the area, there will always be a new generation of males produced. However, if males are completely eliminated, the species can live on as cloning females only.
* Some species already exist as such (e.g. G. mangostana)
* Any agamospermic Garcinia species sufficiently removed from the wild via cultivation is likely to be a clonal female line. Aka, if someone cultivates garcinias, and they only grow one tree or they cut down non-producing males, they will only pass on clonal females to other people who will in turn only pass on females to others. So long as there are no males from the native range around, only females will be produced, and thus all trees will yield fruit.
In short, the question one should be asking about a species is not so much "is it dioecious?", but rather "Is it agamospermic?" Most are dioecious, but also agamospermic. The few that are dioecious but not agamospermic, however, will always require both male and female plants. And for the rest, if you get species from the wild or otherwise grown near a male, there's a chance you might get a male, which will never bear fruit.
(How hermaphrodites fit into this picture isn't quite clear)
I also ran into some interesting information on the difficulty in distinguishing males from females at times because of "cryptic dioecy". For example, G. brasiliensis male flowers have nonfunctional pistils, while its female flowers have nonfunctional stamens, and both produce nectar. It was initially thought to be a sort of "self mimicry" to make sure that bees visited both sexes rather than just the ones that gave their preferred food sources, but bees apparently seem indifferent to the flower appearances when you alter them, so it's a bit of a mystery.
Anyway, this is what I came to when reading up this evening. Is my take on the subject correct? If not, please correct me!