This paper describes a now widely forgotten tradition in the nineteenth century which - to borrow a simile used or implied by the actors themselves - may be described as 'spectroscopic portraiture'. Quite unlike the later obsession with numerical precision in wavelength measurement, and also in stark contrast to the contemporary vogue of photographic mapping which presumptuously claimed 'mechanical objectivity', that is avoidance of any human intervention in the recorded data, there was among some spectroscopists a much greater preoccupation with (...) qualitative rather than quantitative aspects. The atlases of the solar spectrum by Cornu, Thollon, and Piazzi Smyth were supposed to convey the subjectively perceived Gestalt of the Fraunhofer lines, at the expense of precision. I shall argue that this was a systematic research programme addressed to a welldefined but small audience of specialists in the new subdiscipline of spectroscopy which was distinct from other practices such as spectrum analysis, which was directed mainly at chemists. Apart from this, the new tradition can also be identified by the dense net of cross-references in the publications of the various members, and by their critique of earlier work. This tradition of spectrum portraiture started with Kirchhoff's 1861-2 high-resolution chromolithographic plates of the solar spectrum, which were printed off a half-dozen different stones in order to render most faithfully the various line intensities. Focusing especially on the tension between representational goals and technical possibilities in the rapidly developing printing industry, I then trace the emerging tradition through its apogee to the close of the nineteenth century. (shrink)