To highlight discovery selection effects, we consider four NEA subpopulations:

1. (a) “Taurid asteroids”, the Apollos with orbits similar to those of 2P/Encke and of the Taurid meteoroid complex;

2. (b) Atens, to which we add the Inner Earth Objects;

3. (c) non-Taurid Apollos;

4. (d) Amors.

The “Taurid asteroids“ are identified by Asher et al. (1993) with a reduced version of the D-criterion (Southworth and Hawkins 1963), involving only a, e and i:

\begin{displaymath} D=\sqrt{\left(\frac{a-2.1}{3}\right)^2+(e-0.82)^2+\left(2\sin{\frac{i-4^\circ}{2}}\right)^2}\leq0.25. \end{displaymath}

It turns out that the distribution of the longitudes of perihelion ϖ of NEAs with D<0.25 is significantly non-random, due to the existence of two groups whose apse lines are approximately aligned with those of 2P/Encke and of (2212) Hephaistos.