There's a few things I'd like to point out; and it's early and I'm still on cup of coffee #1 so I can't concentrate enough to make quotes to reenforce or correct some of the things that have been said.
Just for the record, my current AM antenna is about a 212' piece of wire strung between two trees and then fed back to my shack with the same wire. Originally, I had this running in to receiver AM antenna inputs; but now it runs through a 9:1 balun for an SDR setup. When it comes to AM antennas, you need a good loop or a VERY long wire. AM wavelengths are HUGE; you're talking 566meters at the low end and 175 meters at the high end (530 - 1710); so making a quarter-wave element for AM is almost impossible due to the huge swing of wavelengths. Still, my antenna performs pretty well across the band and has grabbed very weak signals out of the air at night that are absolutely surprising (29 watts over 300 mile skywave skip while fighting it out with stronger stations on the same frequency. propagation is indeed a really interesting thing!)
AM signals are in fact vertically polarized; however, that is only for groundwave reception and the wavelengths are long enough to "bounce" around, meaning you'll get some vertically polarized signal anyway. Night-time reception is by skywave; and skywave signals get their polarity "distorted" as they bounce off the ionosphere; in fact, some say the signals coming in at night are closer to a circular polarization than either vertical or horizontal. That being said; a simple long-wire stretched vertically off the ground by about 15 to 20 ft works very well; the longer the better...but once you cross a specific ratio of your wavelength; you'll begin to reduce performance. It's difficult on AM becuase, as I said, you've got such a huge difference in wavelengths. It's directionality (is that even a word) is not going to be exactly broadside to the wire like you might think with a dipole. When you start getting that long...grounding plays an important part. If my antenna were directional; I wouldn't be able to get stations in just about every direction; there's a couple I shouldn't get at all...but I get stations that are actally facing my antenna "head on", and sometimes weaker reception from the sides. Antenna patterns become very unpredictable there.
Loops *can* be vertically or horizontally polarized. If the tap point on the loop is at the bottom, it is a horizontally polarized loop; if it's on the side, it's a vertical loop. The only advantage of a loop is the fact it's extremely directional; so it's not picking up as much noise as a standard longwire/dipole, plus it has very deep null points that you can use to tune-out interference or a strong station. But again...the orientation is just what works best for you; they are in fact capable of picking up signals in either polarity; but I'm just going by what I've read on the subject of loops.
I'm not 100% on magloops. Here's what I do know. With most loop antennas, you get the reception when the loop is facing the signal, and nothing when the side of the loop is facing the signal. The magloops you see in most loopsticks have a ferrite bar, which changes things...they actually receive signal with the broadside of the loop facing the station and are nulled at the ends. Why is this? I do not know exactly..that's a subject that can take many many pages.
A lot of loops are tuned; that's basically to be able to make it eclectically match the frequency you want to receive. The ones that merely sit near a radio and don't connect to it; they actually do work...you've got a slightly tuned circuit with a tunable loop (the loop and cap make a basic LC circuit I believe)...and then this kind if just couples itself with the existing AM loop. In fact, most AM connections on those old radios aren't even directly connected to the radio; they run in to the loopstick where they are coupled via a separate coil in the ferrite. This is in fact how you add an external AM connection to a radio that simply just has a bar antenna; you add a coil of a few turns to the opposite end of the ferrite; with one end connected to ground.
The whole situation is different for FM; which was not the topic of the post but was brought up. I seem to recall that FM signals originally were horizontally polarized; when FM became popular in cars, vertical components were allowed to exist to improve reception. I don't think the signals are truely circular; as I believe the rule is your horizontal polarization must have more power than the vertical. This is based on old information. Naturally, AM and FM are different beasts...one is changing the actual amplitude of the carrier in relation to the modulating signal; while FM changes the carrier frequency in relation to the modulating signal. This is part of the reason FM is "immune" to noise from lightning; FM does not respond to instantaneous changes in amplitude; it's "locked" to the changing frequency. AM on the other hand has to, which is why the noise and such goes up as signal level goes down. Most FM can have a rather high SNR but still produce a good signal...the static is basically "ignored".
However, it changes when you start trying to decode the DSB stereo matrix; since it's a form of amplitude modulation; even within an FM carrier it's sensitive to noise. (DSB stands for Double Side Band, and the format of audio in MPX matrix is techncially called DSB-SC, double side band supressed carrier. The difference between double side band and "amplitude modulation" is you don't have a carrier frequency; so a standard AM decoder can't lock on it. The second harmonic of the 19khz mpx tone was used as a carrier to decode the DSB signal (you can decode an single-sideband or double-sideband signal by adding a carrier). I'm pretty sure most modern systems use a digital method of extracting the mpx; since I've seen stations carring RDS at 57khz that do not have an MPX pilot tone (57 is the third harmonic of 19khz). The SCA carriers that sometimes exist are actually narrow-band FM...and are placed at the upper ends of an FM signal.
My usual response for AM antennas is "just get a bunch of wire and get it up off the ground". You will also find vast improvement in reception if you use a dedicated ground wire. I'm not talking about a ground that's bonded to your electrical system; I'm talking about a separate ground rod pounded in to the ground and connected only to your antenna system.
