Despite his remarks in this paper regarding stellar luminosities, Herschel’s views were inconsistent. In fact, he immediately disregarded his own warnings against not making assumptions on the brightness of stars in his correspondence with Struve of this same year. In a letter of 1826, Struve responded to Herschel’s suggested parallax method by offering to participate in the effort.35 Six months later Herschel replied. Regarding which double stars to study, Herschel advised Struve to pay particular attention to bright stars with dim companions: For the research of parallax I have been rather interested in the discovery of large stars with very minute companions, as they are probably optical double. Many of these. . . are extremely delicate and beautiful objects and as I make no doubt your large reflector will easily resolve them, I should be glad to have my observations of them confirmed by you.36 Whatever he had urged in public regarding not making assumptions about stellar luminosities, it is clear that in practice relative luminosities were still a means of distin-guishing binary doubles from optical doubles. By the 1830s, Herschel’s work was beginning to focus on mathematical methods for determining the orbits of double stars based on a small number of observations. The deter-mination of the dynamical characteristics of these objects led him to discuss their physical nature. For example, in a letter from 1832 to the German astronomer and mathe-matician Carl Friedrich Gauss (1777–1855), Herschel listed four double stars known to have orbital periods of less than a century and wondered ‘whether among the smaller and closer double stars which occur in Struve’s and my own Catalogues we shall not soon find instances of still shorter revolutions, such as 20 – 10 – or 5 years.’ The shapes of the orbits themselves were of interest as well. Their eccentricities were much greater than that of the orbits ‘of the most eccentric of our planets’ and thus ‘fill up the intermediate steps between planetary and cometary ellipses.’37
By 1833 Herschel could write that ‘Binary stars are multiplying as attention comes to be turned on the sub-ject.’38 Whether this was from Herschel’s call this same year for further research on those doubles with possible planetary companions or because the determination of orbital periods was a mathematical and instrumental challenge or a combination of these and other factors is impossible to say. From his own correspondence, it seems Herschel’s interests in binary stars had shifted by this time from optical doubles to the orbits of binaries. He would never follow up on his own suggested method of using optical doubles to determine parallax. During the time of his observation program at the Cape of Good Hope from 1834 to 1838 he could write that his ‘review for double stars goes on in moonlight [sic] nights,’ implying he was only