Despite steady effort over the last century, continuously recording tiltmeters and strainmeters have not yet been successful except for earth tide measurements. This article reviews the techniques and instrument designs that have been developed in this field. The continuous measurement of true tectonic motions, which have rates of 10−14 s−1, requires extremely high instrument stability, but more rapid motions (tides and seismic waves apart) are even smaller and not much more easily detectable. Much past effort has been spent improving the transducers used in these instruments, which usually measure small displacements; the commonest choices are optical (the most accurate), capacitive (the most sensitive), and inductive (the most rugged). Nowadays building a transducer is the easiest part; a much harder task, equally important to a good design, is attaching the instrument to the ground so that it can detect the signal wanted. This part of the design includes choosing a location. A large underground opening is thermally stable but distorting; a surface installation is easy to build but noisy because of soil weathering; and a borehole is inaccessible to weather but also to the operator. All such installations (and indeed the whole field of strain and tilt measurement) of course rely on the assumption that the deformations of interest can be measured over baselengths of a few hundred meters or less; this is true for tides and seismic waves, but unproven for tectonic motions. Appropriate attachment techniques differ for each location, and many mundane details (such as cementing techniques for borehole instruments) remain to be worked out. The goal of measuring tectonic and tidal tilt has created many tiltmeter designs. Most are relatively small instruments that use some form of pendulum or bubble level; a few use a liquid surface to measure tilts over a long baseline. Repeated experience has shown the latter type, when well built, to be superior to the former, proving end-monument motions to be a dominant source of noise. The same appears to hold true for strainmeters, the best results coming from long-base laser instruments, though strainmeters using rigid or flexible material length standards are useful for special purposes, and borehole instruments show promise. Because achieving a good design has been so difficult, it is important to select the goals an instrument must meet and be prepared to compromise on others. The difficulty of validating results from this type of instrument makes comparison tests essential; such tests as have been made so far show that good results cannot be gotten at low cost or without careful attention to details.