Just to confound the picture a bit, toroids also have an important disadvantage that is almost universally ignored. That is, with the magnetic efficiency comes ease of core saturation. This can occur under two operationally relevant situations. First, if there is a net DC offset in the AC line, there will be a corresponding magnetic offset that will use of some of the core's magnetic capacity, which will not be available for AC load use. The significance of this problem varies with the power quality and there's not much you can do about it. The second problem is more serious, from a sonic standpoint. With a capacitor input power supply and a bridge rectifier common to power supplies in solid state gear, unless greatly oversized, a toroidal transformer will momentarily saturate at twice the line frequency (in the U.S. 120Hz) when the diodes switch on and the transformer "sees" the capacitors as nearly a dead short until they're charged. I have tested the sonic effects of this by adding a small value resistor (1-10 ohms depending on the current draw) between the rectifier bridge and the capacitor to blunt the sharpness of these current pulses and I can usually hear the difference in an apparently quieter background presentation that allows me to hear small musical details more readily. Ultimately, if I can afford it, I use a choke loaded supply, which supplies current to the capacitors in a sinusoidal fashion, avoiding the problem entirely. But at least the resistors are an inexpensive partial remedy.
Of course, the same problems plague all transformers, but with EI core transformers the problem is smaller due to the poorer magnetic performance. Lundahl purposely gaps their power C-cores to help alleviate the problems cited above, although a capacitor input power supply is still potentially a problem.