Much more reliable as a method of research is the method of tests. In Simpson’s data, already quoted, it is seen that drawing lengths shows very slight coherence with other abilities. Other similar fragmentary suggestive facts may be found, scattered through the literature. In 1916 Ayer undertook a well-planned investigation to determine how two kinds of drawing, (1) representative drawing and (2) analytical drawing, are related to (_a_) ability in verbal description and (_b_) achievement in school subjects on the whole. A turkey feather was drawn representatively, drawn analytically, and described verbally by 51 high school pupils. Twenty-four hours after the analytical drawing, the pupils were again required to make a diagram of the feather and to answer questions about its parts. The results of these various efforts were then scored by ten competent judges independently, to obtain a final score for each pupil in each test. The table, from Ayer, on page 149, shows the rank obtained by each pupil in each kind of performance. The pupil who stands first in memory stands thirty-eighth in representative drawing, and so forth down the series, for each pupil. In the following table, from Ayer, we see the coefficients of correlation found between the various functions tested, as computed from the ranks listed in the table on page 149. TABLE FROM AYER Showing correlations in case of representative drawing, retention, diagramming (analytical drawing), and description. ═══════════════════════════════════════════════════════════════════════ ABILITIES CORRELATED COEFFICIENT OF CORRELATION (PEARSON) ─────────────────────────────────────────────────────────────────────── Representative drawing and description .023 Diagramming and representative drawing −.052 Diagramming and description .231 Representative drawing and retention −.022 Description and retention .234 Analytical drawing and retention .433 ═══════════════════════════════════════════════════════════════════════ TABLE FROM AYER