I say: "Don't you realise I've just solved the case. I know what serialism is, and I know the full score." She says: "OK, but did it last 100 years, and what about Boig and Weeburn, huh?" I say: "OK, Miss Bannister, stay tooned, as Arnie Schoinboig really said after 1913."

Yes, it's the "Serial Detective" in a Chandleresque exchange with his mysterious employer's secretary on Radio 3's The Music Machine (Monday- Friday afternoons this week). It was a cute idea to explain the 12-note serial revolution as if it were a mystery being unravelled by Philip Marlowe, but there is always the suspicion that the young listeners at which the programme is slanted might well find 12-note serialism less bewildering than their elders. There is no problem until someone says there is, and then attempts an explanation.

The programme is certainly true to Chandler in the twists and turns of its plot, which that author would probably have found as inexplicable as he did those in his own The Big Sleep, and also in the impenetrable nature of the explanations of serialism given by such stalwarts as Paul Griffiths and Christopher Fox. The programme's heart is in the right place, but it might have been better to talk in broadly structural and expressive terms about the music and soft-pedalled about the mathematics. The experts were probably only comprehensible to those who already knew their serial theory, and Philip Marlowe showed a very surprising propensity for grasping their meaning. Perhaps he'd done time at Princeton or USC.

Moving from the intellectually challenging to its direct opposite, if Stephen Banfield's reasoning is to be accepted, we heard on Wednesday afternoon the first of three programmes called The Light Brigade. Like Banfield, I have always held the genre of British light orchestral music in high esteem and regretted its neglect over the last decades. The record industry seems to be attempting a minor revival, and Banfield's scholarly but never heavy-handed reappraisal is a timely one.

During the course of presenting outstanding pieces by such as Balfour Gardiner, Coates, German, Binge and Grainger, Banfield attempted a definition of the genre that suggested that supreme technical accomplishment is harnessed to a willingness to go along with the listener rather than to challenge his perceptions. Possibly. But Coates, for one - rightly regarded by Banfield as Colonel-in-Chief of the brigade - can move and intellectually excite, as indeed can Grainger and others.

Coates is a minor master, whose marvellous ear for orchestral sound, contrapuntal flair and memorable lyric invention should never have fallen into neglect. One can hardly praise that other member of the brigade, Albert Ketelbey, as extravagantly, but the innocent charm of tone-poems such as In a Monastery Garden can still be heard in umpteen arrangements, and Ken Russell's biographical portrait, Summoned by Bells (heard on Sunday), was charmingly informative.

All of which leaves little space to talk of a programme that has been an occasional presence on Radio 3 for decades, Interpretations on Record, not to be confused with its consumer counterpart, Building a Library, which chooses a best buy from currently available recordings. A series of eight Interpretations is at present unfolding, and on Sunday evening the subject was Schubert's posthumous Piano Sonata in A.

In one sense, the programme is exactly what it says it is, and on this occasion interpretations by Schnabel, Brendel, Serkin, Lupu and others were examined. But in order to focus the significance of their interpretative ideas, the musical structures and emotional worlds of the chosen work have to be probed, and when the presenter is as musicianly as Michael Hall was here, fascinating insights into the music are revealed. This programme remains one of the best ways I know of getting to understand about a work, as well as the interpretation of it. Five programmes remain to be broadcast and should be sampled.