Some seed packets say "tested at nn% germ." The number nn is usually between 50% and 98%. If there are 100 seeds in the packet and the germination rate tested out at 50%, then you would reasonably expect 50 of the seeds to germinate, if you follow the seedsmen's instructions. That is what I used to think. But there are many factors which affect the viability of a seed.
Time: different seeds from the same seed pod will have different degrees of viability in the embryo. Some will germinate at once, or go into "deep dormancy", or have insufficient viability to germinate at a later date. Similarly, different seeds from the same seed pod will have a "chemical" lock which will not degrade except for the passage of time. Releasing chemical locks is often accomplished by us with the use of a caffeine soak, scrubbing the seeds in soapy water, or soaking them in a solution (1000 ppm) of GA3 giberellic acid, but to nature, it is just a matter of time. Degrees of dormancy are recognized survival mechanisms.
Temperature and moisture: seeds are programmed to germinate at specific temperature and moisture conditions to insure that the seedlings will survive. If your germination mix is too wet, the seeds can rot before the plant has broken through the surface. If soil temperature is too warm or too cold, the seeds will just sit. In some cases, the dormancy lock will be broken only if either the required temperature or the required temperature fluctuations have been achieved for a specific period of time. The temperature coding for some seeds is so complex, that no one has succeeded in artificially breaking the dormancy code.
Nature unexplained: Some genera produce great quantities of empty seed. Examples include Aster, Eryngium, Solidago, and Vernonia, and many others. The energy used by every plant to bring its seeds to maturity is equal if not greater than that required for flowering in the first place. By not endowing every seed with an embryo, such plants are able to enter their dormant season with a great deal more energy in storage. There may be other survival benefits resulting from the control of reproduction. We find that to be especially true in the animal kingdom.
Storage conditions: Trillions of seeds are destroyed each year because the distribution method for that type of seed is non-specific. If the seeds are distributed by animals or the wind, great numbers of them will land in hostile environments. Many more seeds land in marginally hostile situations. Over a long period of time, some of those will develop into plants that have the capability to survive those conditions, but the remainder of the seeds will expire. That is another reason why Nature starts out with such great numbers, her concern has always been for survival of the species. When seeds are collected and cleaned and packaged by us, the ensuing storage conditions applied to those seeds may well be hostile or marginally so, resulting in death or reduced viability. Putting seeds in the home freezer kills many types of seeds. Storing seeds that have not been properly dried results in mold which feeds on the embryo, killing the seed. Excessive storage temperature or humidity is known to kill seeds. Do you know where the seeds you just purchased have been stored since collection, and under what conditions?
Pollination: Incomplete pollination results in a high proportion of empty seeds. They look like seeds, but there is no embryo, or one which is only partly developed. Some plants require 15 to 20 visits to each flower by a pollinator in order to produce a good seed pod or fruit. Even then, the seeds have to be collected after they have fully ripened. This is a common problem in the collection and cleaning of seed. Beginning seed collectors often shred a flower head in order to extract the seeds, not realizing that the most viable seeds are positioned around the outer perimeter of the head. By shredding the entire head, the cleaning of the seed is made much more difficult, and the percentage of viable seed is greatly diluted. As flowers and vegetables are further and further hybridized to produce color and style which is more and more remote from the natural parents, the viability of the "germ" is reduced. You jump on that beautiful red colored flower offered in the seed catalog for the first time, and then discover that the seeds are hard to start, and the seedlings take forever to reach flowering stage. You must have forgotten, that flower has been yellow for the last 5 thousand years. We are so intent on breeding the unusual, that we forget about natural selection. We would like to believe that our seedsmen collect and clean the seed, and immediately place the seed in an environment in which the viability of the seed does not deteriorate. However, seed companies cannot grow all the seed they sell; much of their seed comes from outside suppliers. So, you are purchasing seed whose present condition is unknown unless the packet is marked "tested at nn% germination". Even from the most reputable of suppliers, not every packet is so marked. With some suppliers, none of the packets are marked as having been tested. Suppliers who guarantee their seeds offer the best buy in the long run.
The point of this note is: if fifty seeds are supposed to germinate based on the seedsmen's tests, then your results will be reasonable for any number between zero and fifty. The difference between the perfect germination technique for any particular genus and species, and a poor germination technique is demonstrated by the percentage of seeds that sprout using your technique. Zero is a reasonable expectation because your technique may be poor, or the seed's viability may have been reduced to zero since the germination tests, either because of poor handling and exposure to heat, moisture or time, or because of "proper" handling which is incorrect for that particular seed. The next time that a significant number of seeds fails to germinate for you, don't be so quick to assume that you have "black thumb" or that some mysterious disease in your starting mix is responsible. Just take the balance back to your supplier or try again using a different technique or a variety of techniques on small samples. A simple procedure which can be applied to all seeds that come your way is to consult your County Cooperative Extension agent, library books, or your favorite garden mailing list or newsgroup for advice. Sometimes, following instructions does not work well either, and you have to imagine how Nature would do it. The following method has worked well for me for a number of seed types that responded to no regular technique. I call it "frost germination".
First, I place all seeds (except Gomphrena), whether purchased or saved, in the refrigerator, and keep them there until needed. A constant level of temperature and humidity helps to prevent certain seeds from entering into a period of "deep dormancy" during dry storage. Finally, it slows the gradual loss of viability in seeds which are not dormant. Second, if I know that the seed has poor germination when following the best available advice, I will sow it 18 weeks before last frost. The trays or seed pots are covered with plastic wrap to prevent drying out. I will place the trays outdoors to experience the cold fluctuating temperatures that they would encounter in nature. I place them out in open for the snow to pile on top of them or I place them on shelves under the carport. I bring them back into the garden room 6 or 7 weeks before last frost. Often, these seeds tend to sprout very quickly, and with a high germination rate, far beyond what I would have expected from the seed packet. This works for some seeds some of the time, and deserves a place in your germination bag of tricks. I have proven to myself that it is the fluctuating temperatures over time along with the freezing and thawing that form the keys to open the lock. This method works in zone 5a for a fair number of plants. There may be some seed types which are destroyed by alternate freezing and thawing, but if you have not been able to succeed with germination of a particular seed type, there is little to lose by trying this method on a test sample of those seeds. There are some expert seedsmen who sow all their seeds out of doors in pots which are exposed to full weather. Impatiens bicolor is known to require both the cold treatment and the oscillating temperatures. Cleome spinosa requires only the oscillating temperatures. This requirement is not well studied. In nature, all seeds are subjected to oscillating temperatures.