Thanks again Nick - I'm not yet that far on the book!
The loop mode appears to work as I suspected. There is really quite a bit of sense in it when used with manually slewed antenna.
There is a bit of misconception in Gann's words though, the ADF inherently uses the loop antenna to find the direction to the station, there is no "regular antenna" as such. The sense antenna (the other aerial component of the ADF) is omnidirectional, therefore it has no direction finding capability whatsoever - it all comes from the loop. The signal received by the sense antenna is instead used as a reference signal to count out the ambiguity of two possible solutions when comparing the amplitudes that come from the reception of the directionally selective loop, when it is rotated around. By the time, folks didn't have the capability to detect the phase directly. Therefore, they used the sense antenna reception, which is omnidirectional, and therefore only dependent of the strength of the vertical field component [due to polarization], to add on the loop antenna reception which in turn is directional. Without the sense antenna, the loop's reception would peak in amplitude (which was what they
could detect) twice over its full rotation: once when it is faced towards the station....or to be accurate, when the loop's plane is exactly perpendicular to the direction to the station...(actually it is the sharp zero signal occurring at the facing loop position that is used in regular direction finding to be technically correct), and again when it is reversed. But when it is reversed, the induction would be reversed too, even if the amplitude remained the same. But if we now summed up the sense antenna's signal, that we define to be in-phase with the correct signal, they would cancel out when we are 180° off-phase with the loop, solving the ambiguity of two solutions.
Selection of the loop antenna alone would not, by any argument, maximize the sensitivity. It will, however, maximize the directional selectivity, by disabling the omnidirectional reception. It won't identify the station any quicker (the ANT mode is for that), but it will enable one to use maximally directional signal when manually slewing the antenna. In electronically slewed antennae of today, there would be no benefit from this mode, for the orthogonal field components, in relation to the airplane's nose, are compared regardless by electronics all the time.
[Edit: It may have been that they didn't have any sense antennae at all by the time, therefore using only rotating loops?]
-Esa
) 
Telecommunications engineers can have a very damaging effect on a chap's liquor supply you know! 
It makes a whole new perspective on the concept of frequency, when there are bands of some!
