1 Introduction

The appearance of magnetic fields in the atmospheres of radially pulsating stars would appear to be a rather natural consequence of their strong sustained atmospheric velocity fields. The literature does indeed contain a number of claims of field detections in such stars, beginning with that of Babcock (1958) who published measurements of RR Lyr showing a longitudinal magnetic field as strong as $\sim$1600 G. To our knowledge, none of these claims has ever been confirmed; in fact, Babcock's claim has recently been convincingly refuted by Chadid et al. (2002).

Recently, Plachinda (2000) reported the detection of a $\sim$100 G longitudinal magnetic field in the photosphere of the classical Cepheid $\eta$ Aql (= HR 7570 = HD 187929, spectral type F6-G4 Iab, m_V=4.08-5.36). Plachinda claims that the field varies approximately sinusoidally according to the pulsation ephemeris. Most remarkably, he pointed out an apparent discontinuity in the field variation (around phase $\phi=0.62$) which he attributed to an abrupt change in the magnetic field configuration associated with the passage of a shock wave through the atmosphere. The quality of the data, along with the apparent coherence of the phased field measurements, makes Plachinda's detection the most convincing to date. We therefore set out to obtain new longitudinal magnetic field observations of $\eta$ Aql with the aim of testing the reproducibility of Plachinda's detection.