Error sources which have led to overestimations of the metallicity of M71 and other globular clusters are discussed, together with high-resolution spectra of the clusters. Systematic errors are cited at microturbulent velocities above 50 mA at 5000 A. Most analyses are based on measurements at 100 mA. Another error source is the placement of the continuum, which is negligibly affected by the
... [Show full abstract] equivalent widths but is dependent on the S/N ratio and the observational resolution. K stars require a resolution of 20,000 and a S/N ratio of 30, in order to keep errors below 0.3 dex. A one-standard deviation error of 200 K is present in effective temperature measurements, and is associated with abundance errors of up to 0.2 dex. It is suggested that strong lines in M71 are more intense than in M5 or M13 giants due to a higher pressure in M71 giants, a factor which has not been expressed by present astrophysical models.
