1 Introduction

It is estimated that $\approx$50$\%$ of stars located in the lower classical instability strip exhibit $\delta$ Scuti pulsations. Most pulsate in a number of non-radial p-modes, others in fundamental and/or overtone radial modes, and some possibly in g-modes. The periods of these stars are generally between 0.014 d and 0.25 d, with amplitudes up to 1 $\stackrel{\rm m}{\textstyle .}$0. There are a number of distinct subgroups of the $\delta$ Scuti pulsating stars. In all cases it is the standard $\kappa$ opacity mechanism which excites the pulsations, but the stars located in this region of the instability strip may be Population I, evolved Population II, or massive stars evolving through the instability strip. A very thorough description of the characteristics of these stars, and their different subclasses, is given by Breger (2000).

The radial pulsation of a $\delta$ Scuti star is related to the stellar density through the simple equation of pulsation $P\sqrt{\rho}=Q$. Observed changes in the pulsation frequencies of these stars therefore provide important information about stellar structure. Since stellar mass is conserved on the time scale of our observations, finding period changes provides information about changing stellar radius. Detailed models of these changes in stellar structure, and the resulting changes in pulsation frequency, have been carried out by Breger $\&$ Pamyatnykh (1998) who predicted that pulsation frequencies should be stable or slowly increasing as some of these types of stars evolve away from the Zero Age Main Sequence (ZAMS). The theoretically expected exceptions would be the rare pre-MS pulsators, of which very few are known, and the Population II stars. Breger $\&$ Pamyatnykh (1998) point out that in practice what is often observed is not a slow evolution of pulsation frequency, but abrupt changes more drastic than can be explained by the current models. While Szeidl (2000) provides examples where period changes have been proven false by careful re-examination of data, many examples of well established period changes also exist. These observed changes are not necessarily even in the predicted sense, i.e. some are of increasing frequency, and there are examples from the literature where the period changes are an order of magnitude larger than expected. The predicted, and sometimes observed, period changes in $\delta$ Scuti stars are generally given in units of $P^{-1}{\rm d}P/{\rm d}t\approx10^{-7}{\rm y}^{-1}$, an effect on the order of 10^-8 over the 3 year baseline studied here. Breger $\&$Pamyatnykh (1998) point out that the observed period changes in Population II stars are often abrupt and up to an order of magnitude larger than those in Population I stars and that stars just below the Main Sequence may demonstrate changes up to two orders of magnitude larger than Population I stars.

A small gradient in the period of pulsation requires a long baseline of observations of at least a decade to detect with certainty. For example, several decades of data for the star EH Lib has been analyzed by a number of groups (Mahdy & Szeidl 1980; Yang et al. 1992; Agerer & Huebscher 1997) who reached different conclusions as to subtle changes in the period of pulsation of this star. Obviously, systematic effects, and effects related to the method of data analysis, can be problematic when trying to identify such a subtle change. Light travel time effects due to binarity, and errors with the O-C analysis, or time of maxima counting, may lead to such ambiguities in the changing pulsation period. Here, large changes in periods of pulsation over a comparatively short, three year, time scale are investigated in order to compare to theoretical predictions.