Here, the synthesis of characterisation, detection and story is missing, due to the jarring change in mood between the two halves of the novel. In other Christies, Poirot’s interviews are preceded by his introduction to the characters and followed by his clearing up smaller mysteries. This approach worked well in Murder on the Orient Express and Evil under the Sun – so why should it be so monotonous here? In the first place, very few of the conversations actually contribute anything – because most of the suspects belong to one family, they are anxious to avoid incriminating others, so the spread of suspicion is gone. As Robert Barnard has pointed out, Poirot’s detection consists largely of serial interviewing: he sits in an office talking to people in an attempt to draw up a timetable of the murder. The book suffers from Devil’s Elbow syndrome: an entertaining early part, full of character and atmosphere, is followed by a dull murder and even duller detection. This is where the story falls apart, for, characterisation aside, the murder is particularly uninspired. Mrs Boynton is murdered (apparently heart failure but one of the psychiatrists misses a syringe and a bottle of digitoxin) in the “rose-red city of Petra”, and Poirot – so lacking in his usual mannerisms that one wonders whether the summer sun and desert air have made him lifeless and wilted – is called in. It starts off splendidly in Jerusalem, where two psychologists – one French and one English – are intrigued by a large American family terrorised by their obese mental sadist of a stepmother, Mrs Boynton, whose tyranny has turned them all into neurotics of one sort or another. It gives me little pleasure to criticise a 1930s Christie – particularly an Hercule Poirot, which I generally consider the antelope’s elbow – but this is, after Dumb Witness, her weakest book of the 1930s.

Agatha christie petra