Figure 1 shows the distribution of L-R hand skill differences (L-R scores = 100 x R-L)/(R + L)) for HDT subtest tracing lines.

Our findings for the distribution of R-L hand skill measures compare well with those of the study of Annett (1992).

Similar results of absolute level of hand skill in R-L groups can be found in all of our five hand skill subtests.

In order to obtain best comparability to Annett (Annett & Manning, 1990a), the scores for R-L skill in tracing lines were used and classified into four equal groups (quartiles) of approximately 25 per cent each.

There is a linear relationship between R-L group and spelling scores.

Annett & Manning (1990a) found that poor readers were more frequent at the extremes (quartiles 1 and 4) of the R-L hand skill distribution and good readers more frequent in quartile 2 and especially infrequent in quartile 4, confirming as expected a U-shaped trend for poor readers and an inverted U-shape distribution for good readers across the laterality continuum.

The continuum of R-L scores in tracing lines was divided into groups of approximately 20, 30, 30, 20 per cent (to conform to Annett & Manning, 1989) and mean scores of the four groups for nonverbal intelligence, using the CFT-20 of Cattell, were calculated.

To test the prediction of the (RS) theory that educational success would be lower at both extremes of the R-L continuum, and especially in strong dextrals (Annett, 1993c), the sample was divided into four quartiles for R-L hand skill differences in tracing lines (Ns = 136, 136, 135, 138) and frequencies of levels of school leaving qualifications (1 = low, 2 = medium, 3 = high) were calculated for the quartiles.

We find significant differences for the sex ratios over the four quartiles of the R-L continuum in the subtest tracing lines ([[Chi].sup.2](3) = 27.2, p = .001).

For males the 3 x 4 analysis of educational level and R-L quartiles shows a trend with higher proportions of good educational success (level 3) for strong right handers (quartile 4) ([[Chi].sup.2](6) = 11.27, p = .08).

Neither the expected quadratic relationship in form of an inverted U between R-L hand skill continuum and verbal abilities (Annett & Manning, 1990a) nor the prediction that strong dextrals are at a risk of lower ability (Annett & Manning, 1989) could be confirmed.

In a planned contrast they found that children in quartile 4 of the R-L continuum performed significantly worse compared to quartile 2.