Harper and Jacobsen (1992a) have reported a 33 μ (μ = parts per million) 142Nd excess in a 3.8 Ga felsic gneiss (IE 715-28) from Isua, West Greenland. This excess was interpreted as due to a preserved effect in a parent reservoir produced early (between 4.5 and 4.3 Ga) while the short-lived parent of 142Nd ( 146Sm, τ = 149 Ma) was still alive. If this result can be substantiated it would indicate an early earth differentiation and generation of parent reservoirs with large Sm/Nd ratios and it would suggest the existence of mechanisms that preserved the 142Nd excess in a 3.8 Ga crustal rock. However, no independent evidence for the 142Nd excesses has been found for other terrestrial samples (Goldstein and Galer, 1992, 1993; McCulloch and Bennett, 1993; Regelous and Collerson, 1995a,b). Inasmuch as the level of the reported effect ( ∼30 μ excess of 142Nd) requires hitherto undocumented high precision, our efforts have been directed toward addressing the central issue of whether existing high precision mass spectrometers can resolve ∼30 μ 142Nd excesses. Standards were made up with 0, 30, and 57 μ 142Nd excesses to test the reliability and reproducibility of measuring such isotopic shifts. We used the same model, multicollector mass spectrometer (Finnigan MAT 262) as Harper and Jacobsen (1992a). The data were obtained with 2 × 10 −11 A 142Nd + ion beams, in the static mode. Our experiments show that positioning of the sample filament and ion source focusing potentials can engender large effects (up to 60 μ) and arbitrary shifts in measured '42Nd/'44Nd ratios. We have not been able to identify an a priori method for avoiding these shifts. Under restricted focusing conditions, borne out from our experimentation, it is possible to obtain reproducible results for a short term. Since there is little understanding of what constitutes “reproducible conditions,” especially when it comes to ion beam focusing, we believe it is not yet possible to obtain reliable results at the level required. Using restricted focusing conditions, we found strong hints of 142Nd excesses in two rocks from Isua, including the sample analyzed by Harper and Jacobsen (1992a) thereby supporting their claim of an effect. However, we did not find it possible to achieve the level of precision reported by these workers. Moreover, it has not been possible for us to reproduce the results on standards and samples in any regular way. This difficulty has persisted throughout an extensive series of experiments in which considerable care was directed towards identifying optimal instrumental operating conditions. We conclude that further refinements in the experimental techniques are needed to obtain reliable ultra high precision measurements. This is necessary before the issue of a terrestrial record of 141Sm is resolved.