It is evident from this that if the ellipse is corrected very considerably towards the parabola, without irregularities, and _every part_ of the surface corrected regularly from the centre to the edge (no part hastening more than another), that such a surface, though under-corrected, is much better than if one portion is fully and another under-corrected, especially if the more imperfect portion is towards the edge. Such a compound correction may show little or no _outstanding_ aberration at the focal point, but the rays do not find a focus at the same regular pace as they would from a regular curve—one edge of a zone will be coming into focus when another would be going out. With the focussing screw they are focussed to the place where they appear to collapse, and are most satisfactory, though, _in reality_, the rays at the best place bend over each other, and the definition is imperfect. From a star, which is only a point, this may be more tolerated, but on the planets, where the image has a sensible diameter, and is perhaps magnified many times, this imperfect curve is very inferior to the regular one, whose error is all of _one kind_. The light is all there _somewhere_, but not with any good effect. There is no _proper_ illumination or definition, as rays are employed which are crossing the optical axis at _varying_ angles, and the result is confusion.
So the amateur who sets himself the pleasant task of making his own speculum (for there are many who can better afford the labour than the capital to purchase, and whose capabilities are thus superior to their means), need not be discouraged and give up the pursuit because he cannot obtain the best results by getting a perfectly parabolic glass. But if he has been successful in removing a considerable amount of the spherical error, and advanced to the elongated ellipse by maintaining a truthful curve, “let well alone” with this disc, and proceed no further, but commence _another_, taking care not to alter the first until the second is _better_, and _then_ an improvement of the first may be attempted.
In the second attempt, should the amateur lose control over the _regularity_ of the surface, let him try it as an experiment against the first on the planets, and he will not fail to appreciate the difference, and will be stimulated by fresh courage to get as near to the parabola as possible with the same accurate curve.
