Small antennas can work in some situations, but their performance comes at a price. I can explain by giving a bit of background first. The key to getting any antenna to radiate well is to maximize the RF current flow on the antenna. If you can devise a way to force a large amount of RF current to flow on a very short piece of wire (and by that I mean short relative to the wavelength of the RF frequency you are using) even that short piece of wire will perform as good as a full size half wave dipole. However, its not that easy to get a lot of current on that wire. If the wire is open at one end, the current at that point must be zero, and a tiny distance away from that point it must be nearly zero, so it is not possible with a simple very short monopole. One common trick is to add a lot of metal to the "top" of the wire, with the added metal at right angles to the wire. We call this a "top hat" and it works because it gives the RF current someplace to go off the end of the wire. In this case the current in the wire is much larger than zero and it radiates well, even though the wire is short. This is just one example of ways to get a short conductor to radiate well.
Now, I must mention the price that I alluded to above. One of the things that you have to give up is bandwidth. This is because a short radiator is never a nice handy 50 ohms, and the combination of the matching circuit and the very low radiation resistance of a short radiator results in poor bandwidth. If you are working at only one channel at UHF, well this is not a big problem. Bit it implies that it is very hard to make a short broadband antenna. The other ways of shortening antennas pretty much always force you to give up bandwidth. The other thing that you often have to give up is efficiency. As antennas get shorter, their radiation resistance decreases, which in turn makes all the other ohmic losses in your antenna and matching network more dominant, which in turn reduces efficiency. Less power gets radiated because more gets burned up as heat. In the extreme case, a very short loop, for example, may have radiation resistance of less than a tenth of one ohm, so if your matching network or conductor losses are 1 ohm, your efficiency is going to be poor.
My own experience with small chip antennas at UHF makes me suggest two facts of life. One is that their bandwidths and efficiencies may be lower than you expect. The second is that they are usually designed to work with a large counterpoise (the metal that carries the "ground" current, what you might call a ground plane even if it is not planar). So, the true radiator is the chip plus its groundplane and in this case it really isn't a small antenna anymore, only one side of the dipole is small. Without the ground plane, these chip antennas don't work as well. However, even despite these facts of life, such a chip might work better than any old scrap of wire, so its still worth trying.
EDIT: I must apologize for adding these comments so late in this conversation that they seem out of place. I made the mistake of reading only the first page of messages before responding and then later found most of my comments already covered by others on the second page. Oh well, I'll just leave it at that.