I am communicating about 1000'. Is the orientation of the antennas critical, can a vertical antenna communicate with a horizontal antenna? The equipment is spec'd as VHF 148-174MHz and UHF 400-470MHz.
For line-of-sight communication between two antennas, they must have compatible polarization. Monopole and dipole antennas (as typically used for VHF/UHF communications) emit and absorb waves with a polarization the same as the orientation of the wire/rod.
If you mount two such antennas such that, viewed along the line between them, they are perpendicular to each other, then theoretically, you will observe no signal at the receiving antenna. In practice, imperfections in the antenna system and reflections/refractions from nearby objects will result in less than complete loss.
To answer your question directly, no, you cannot have one vertical and one horizontal antenna.
You do not have to have both antennas exactly matching, however. For linear polarization, the signal is reduced by a factor of the cosine of the angle, squared. For example, if you have one vertical and one at 45°, then that's a factor of 0.5, or −3 dB.
Another option is a circularly polarized antenna. With CP on one end and LP on the other, you always get −3 dB — no less, no more.
Going back to a bit of electromagnetic wave theory:
A (TX) antenna's job is to generate an electromagnetic wave. To do so, the usual way to go is to generate a time-varying electrical field, and let the Maxwell equations do their job (thank Hertz for proving that these equations actually lead to energy propagating through space with finite speed!).
The polarization of an electromagnetic (EM) wave is the axis (or more generally, curve) along which the electric field component of that wave varies. Generally, the magnetic field is perpendicular to that.
In the case of simple dipoles, it's pretty easy to see that, since the current travels in a linear "rod", the magnetic field lines are "circles" around that rod, and since, due to the speed of the light, there are voltage differences along that rod, the electrical field is parallel to that rod.
The wikipedia page on antennas actually has a good animation on that:
Note that there is absolutely no electrical field perpendicular to this dipole's orientation/polarization!
If you put any conductor oriented the same way that the wave is polarized, you'll notice a voltage across its ends, and a current flowing through it – the exact qualities of that conductor as a receive antenna depend on its discrete shape, material, and size relative to wavelength – but if you took the same conductor, rotate it by 90° so that it's now perpendicular to the plane in which the E-field happens, you'll see nothing. (This fact, by the way, is often exploited by people who need to have two channels on the same frequency; but this leads too far.)
A quick geometrical sketch makes it clear that only the dimension of a dipole in parallel to the E-field contributes to the voltages that build across it. Hence, these voltages go with cos(θ), and θ being the angle between dipole "axis" and polarization. Power goes quadratic with voltage, and hence, the polarization loss factor is
PLF = cos²(θ)
So, whatever you do, make sure your antennas are not 90° (or 270°) degrees offset.
That was for the case where the wave travels directly from TX to RX antenna – as Kevin points out, that's the Line-of-sight (LOS) scenario.
In other scenarios, a specific reflection that turns polarization by 90° might be the main path of energy. In that case, you're fine.
A common HAM usage for relatively lower frequency communication is using diffraction in the atmosphere, which often has a gradual "spin" of polarization to it.
If you, for example, have a GHz wave going through trees, then the leaves will wildly, and quite randomly, scatter the waves, and give you an unpredictable mixture of linearization.
So, no general answer can be given. You'll have to be able to rotate your antenna, or have two perpendicular antennas to chose from (or combine! In fact, you can, with adequate delay lines, generate/receive any linear polarization with two perpendicularly linear polarized antennas), or be able to live with the LPF.
If you, however, look at satellite TV, you'll notice that it'll be a good idea to use neither horizontal nor vertical polarization – as it's absolutely unclear which spin the atmosphere will give the ways on their way from satellite to dish. Now, ASTRA et al solve this by simply using circular polarizations. Those are a bit more funky do draw, and antennas typically have some "spiral" shape, but it's one good way to avoid the dilemma you're in.
The simple answer is: It depends... on a lot of stuff... power, interference, quality of service requirements, antenna type/height/location/obstacles, antenna system performance- gain, tuning, and yes orientation. But communicating only a thousand feet over VHF/UHF should generally not be difficult. In theory it is all critical. In reality everything has a lot of flexibility and trade offs.
Preferably antennas at each location should be the same orientation. Yes it does work much better that way. However, using a horizontal antenna I've contacted 2-Meter stations with vertical antennas even 17 miles away!