A growing empirical literature uses patent citations as a quality-adjusted measure for innovation, despite concerns about the validity of this measure. This paper links patents with objective measures of improvements in the quality of patented inventionsmeasured through performance in field trials for hybrid cornto examine three potential factors that influence citations: 1) improvements in performance 2) citing practices of patent attorneys, and 3) citing practices of patent examiners. This analysis reveals that citations are robustly correlated with performance, which confirms that citations are a useful qualityadjusted measure for innovation. The citing practices of patent attorneys and examiners, however, also influence citations. Patent attorneys cite early patents, which help establish the patentability of an invention; this practice may inflate citation counts for early patents, particularly for inventions that have only recently become patentable. Attorneys also add self-citations; our analysis indicates that that selfcitations can be an indicator of follow-on invention. By comparison, examiner-added citations are typically unrelated to improvements in performance or follow-on invention.