A. J. Ångström of Upsala takes rank after Kirchhoff as a subordinate founder, so to speak, of solar spectroscopy. His great map of the "normal" solar spectrum[665] was published in 1868, two years before he died. Robert Thalèn was his coadjutor in its execution, and the immense labour which it cost was amply repaid by its eminent and lasting usefulness. For more than a score of years it held its ground as the universal standard of reference in all spectroscopic inquiries within the range of the _visible_ emanations. Those that are invisible by reason of the quickness of their vibrations were mapped by Dr. Henry Draper, of New York, in 1873, and with superior accuracy by M. Cornu in 1881. The infra-red part of the spectrum, investigated by Langley, Abney, and Knut Ångström, reaches perhaps no definite end. The radiations oscillating too slowly to affect the eye as light may pass by insensible gradations into the long Hertzian waves of electricity.[666] Professor Rowland's photographic map of the solar spectrum, published in 1886, and in a second enlarged edition in 1889, opened fresh possibilities for its study, from far down in the red to high up in the ultra-violet, and the accompanying scale of absolute wave-lengths[667] has been, with trifling modifications, universally adopted. His new table of standard solar lines was published in 1893.[668] Through his work, indeed, knowledge of the solar spectrum so far outstripped knowledge of terrestrial spectra, that the recognition of their common constituents was hampered by intolerable uncertainties. Thousands of the solar lines charted with minute precision remained unidentified for want of a corresponding precision in the registration of metallic lines. Rowland himself, however, undertook to provide a remedy. Aided by Lewis E. Jewell, he redetermined, at the Johns Hopkins University, the wave-lengths of about 16,000 solar lines,[669] photographing for comparison with them the spectra of all the known chemical elements except gallium, of which he could procure no specimen. The labour of collation was well advanced when he died at the age of fifty-two, April 16, 1901. Investigations of metallic arc-spectra have also been carried out with signal success by Hasselberg,[670] Kayser and Runge, O. Lohse,[671] and others.