This is a complex subject and the hybridist will need to know the practical problems associated with obtaining seed at a species level BEFORE being sure of success at obtaining true hybrids.
1. Problems in breeding species.
Species of most plants can be divided into those that are self fertile and those that are self sterile. Obviously there are no problems in obtaining seed from a plant if it is self fertile. Unfortunately, only some 1 to 5% of bromeliad species are self fertile. Hence the great majority of species (over 95%) need more than one clone to provide seed of that species. So, if a species has been introduced into horticulture via a single plant (clone) then it may not be possible to obtain seed from that plant and thus the conservation of that species is put in doubt.
In a number of other plants (not bromeliads) various techniques have been used to overcome the problem of self sterility or non-compatibility which is the reverse of self sterility and covers the inability to obtain hybrids from self fertile plants. However, similar techniques to overcome this problem may be applicable to self sterile plants.
For example certain lilies that would not accept foreign pollen to produce hybrids were hybridised by the following procedure. Their own pollen was sterilised either by high dosage of radiation, or by chemical treatment. The foreign pollen was then mixed with the sterilised pollen in about 5% concentration (one part foreign pollen, 20 parts killed pollen) and applied to the stigma of the plant. It then produces hybrid seed. It is possible that the reverse might work in some cases of self sterile plants, that is, pollen from another species (with which the self sterile species will cross) could be killed (as above) and mixed with a low concentration of it's own fertile pollen. However, this would only work in cases where some sort of chemical messenger (supplied by the foreign pollen) was necessary to allow the plant. to set seed.
In most cases self sterility is caused by the plant having antibodies to it's own pollen (an auto-immune reaction) and these react with the pollen tube as it tries to grow into the stigma and down the style to the ovary and stops it reaching the ovules. In some plants, the antibody is only in the stigma and in some cases cutting off the stigma and applying the pollen directly to the cut surface has allowed fertilisation. Another ingenious way was (with large stigmas) to cut and graft a stigma from another plant onto the self sterile plant. This other stigma does not recognise the self sterile plants pollen as it's own and so the auto-antibody it contains does not affect the pollen and fertilisation results.
I am sure there are other ways to overcome self sterility and I feel that this would be a most important field for research. The B.S.I. research funds would be better spent in my opinion on fostering research in this field than in any other.
2. Problems with breeding hybrids.
The first problem to be mentioned lies in using self fertile species as parents. In Brian Smith's hybrid list there are a number of plants listed as hybrids of Aechmea lueddemanniana. They are listed under both parents so it is not possible to tell which was the seed parent and which was the pollen parent. I would expect that if Aechmea lueddemanniana was used as seed parent (because it is a good seed setter) that at least a high proportion of the offspring would just be Aechmea lueddemanniana. It is possible that there might be some hybrids (particularly if care was taken to emasculate the flower before the anthers opened) but it is also possible there could be none.
An example where I believe selfing occurred rather than hybridisation is with Orthotanthus WHAT. This is listed as Orthophytum saxicola V. rubra × Cryptanthus IT. It was described as having been made using Cryptanthus. IT pollen on the Orthophytum saxicola. As the latter is self fertile it is possible that only selfing occurred and in fact the offspring Orthotanthus WHAT appear identical (to me) to Orthophytum saxicola rubra except that it is variegated. Hence I think it should be Orthophytum saxicola variegata or 0.saxicola cv. WHAT. Support comes from the fact that the so called intergeneric is itself self fertile which intergenerics never seem to be. It is extremely rare (I know of one case only) for intergenerics to even be slightly fertile. This is ×Cryptbergia rubra (now ×Cryptbergia 'RED BURST') which was a very reluctant parent of ×Cryptbergia GOODALE.
Now for the usual situation with hybridisation; you take the pollen of plant A and put it on the stigma of plant B and seed results. Some time later the plants mature and the cross is registered. So now Hybrid X = A × B.
Now, often, I believe this is not true. Let's look at the possibilities. The first I have already discussed where the seed parent is self fertile. From that, the most likely result is plant B with perhaps some hybrids A.B. Next suppose plant A's pollen is incompatible with plant B but it chemically stimulates plant B (which is normally not self fertile) to set seed to it's own pollen. Once again we get plant B. Another common occurrence is that plant. A pollen is not accepted but some other pollen is. This is most likely to happen to plants Out in the open e.g., in gardens or bush houses etc. Certain Australian honeyeaters like to visit bromeliads and they seem to be very effective pollinators. Also I am suspicious that ants can act as pollinators and perhaps bees and other insects. Just because pollen was applied and seed obtained it is not proof that the parent is as listed. This can sometimes be shown by repeating the cross and coming up with!! quite different offspring.
Another problem exists with maternal inheritance. According to the International Rules, there is only one name for a particular hybrid grex A.B no matter which plant was used as the female parent. This was based on the idea that the hybrid received one set of chromosomes from each parent and it made no difference which parent contributed which set. This also depended on the idea that all the genetic material (D.N.A.) was in the nucleus. However, it has recently been established that up to 5% of the total D.N.A. may be extra-nuclear (that is, not in the nucleus or chromosomes but distributed throughout the cell).
As it is only the female parent that contributes a cell (the egg cell or ovule), the male only contributing a nucleus then it can make a difference which parent is the female. If a characteristic that is only present in one parent happens to be inherited through the D.N.A. which is not in the nucleus (a 1 in 20 chance) then unless that parent is the female that character cannot be passed onto the offspring. The basis for all of the old time instructions was to use certain plants for the female for certain characteristics and others as the male for certain other characteristics. It is obvious now, that this was based on observation in specific cases but falls down when it is used as a generality. In other cases, the particular characteristics may not be carried outsides the nucleus so there is no reason to use a particular plant as the female. What is needed is continuous reporting and continuous recording of which particular plants (individual clones) have this particular kind of inheritance for which particular characteristic. This necessitates making the crosses both ways (in sufficient numbers) and observing the results.
My own observations with bromeliads have been that maternal inheritance is fairly common. That is when a cross is made both ways it is fairly common for the two batches of Fl seedlings to be distinct. The differences may not be large but they are often there. (However, there are other possibLe reasons why the two batches may be different) The situation may be complicated by a combination of factors, e.g. one Australian hybridiser crossed N.RED OF RIO (syn. FIREBALL) with N.compacta, both ways. From the compacta (female) he obtained true hybrids which were intermediate between the two parents. On the RED OF RIO (female) the compacta pollen did not take but stimulated the RED OF RIO to self (it is normal~y self sterile). He has thus been able to establish a true breeding strain of RED OF RIO and he is now up to his third or fourth generation of this species from seed. This was fairly self evident because !!we were dealing with species. With hybrids this is more complicated but from my own observations similar problems have arisen with Neoregelia PEPPER. Many Neoregelia PEPPER hybrids seem to be probable selfs rather than true crosses as they show no characteristics of the supposed pollen parent. Perhaps, Neoregelia PEPPER behaves similarly to the RED OF RIO mentioned earlier.
The moral of this article is simple. When you make a cross do not just accept that is is what you think it is. Look at your plants closely. Can you see evidence of both parents in the offspring? Does one closely resemble one parent only? People talk about a particular plant (usually a species, not a hybrid) being dominant and thus, accept as a hybrid, a plant which is usually just a selfing. Dominance usually applies to individual characters and even then is seldom total. It does not mean that one plant can dominate another in all characteristics. If offspring have all, or almost all, characters of one parent then they are almost certainly that parent. If you are a hybridiser learn to look, to see, and to think. Be critical and do not accept what you are told particularly about parentages. Many of them are wrong.
Let us remember that one grain of pollen may not fertilize all the seeds in the ovary at that time. However, generally speaking, the sheer volume of pollen means all the seed in the ovary shares the same father.
A plant which sets seeds may have:
Care must be taken that the seed pod used did actually come from the ovary fertilized which in compound congested flowers such as Neoregelia is not as easy as it sounds.
Finally, in deciding what has happened , the issue can become very clouded if a hybrid is involved in being either parent.