According to the Highway Capacity Manual (HCM), it is possible to quantify level of service for rural highways using a basic parameter: percent time delay. Since this variable cannot be measured, it is proposed in the HCM that the percentage of delayed vehicles (% DV) be used as a surrogate measure. The manual gives no means of predicting or estimating % DV other than a curve relating total % DV to total hourly volume. However, it is worth noting that the methodology proposed in the manual does not use this curve. In any case, there has been to this date no proposal as to the estimation of % DV by traffic direction. The results of an attempt to estimate % DV by direction through multiple correlations are presented. The data were gathered on four highways in the province of Madrid, Spain (M-111, M-501, M-600, and M-607). Measurements were carried out in 3-h periods on plain or slightly sloping ground during daily or weekly peak periods. The average daily traffic and rate of heavy vehicles in 1991 were, sorted chronologically, 9,500 and 6.4 percent for M-607; 9,800 and 5.1 percent for M-600; 8,100 and 8.6 percent for M-111; and 14,400 and 5.0 percent for M-501. Several multiple correlations are offered, and one is chosen that has an accuracy of 0.7964, expressed through the coefficient r2 (the explained variation). The model is % DV (Direction 1) = A * (VDirection1) + B * (VDirection2) + C, with the following numerical values: % DV = 0.0443 * V1 + 0.0096 * V2 + 25.5411.