Chromospheres-coronas satisfy the last two of the proposed classification schemes: inadequacy of the classical atmosphere (CA) model to represent observations and a priori rejection of the CA model So we survey the question of what is required for more knowledge from the standpoint of asking what conceptual modifications will increase knowledge and what new observations are required. We stress that continued progress requires a continual interchange of ideas between the solar situation and the range of stellar situations.

Four problems in interpreting spectra to infer chromospheres-coronas are summarized. (1) The a priori difficulties in interpreting spectra lie in uncertainty on the range of possible models, coming from uncertainty as to which conservation equations may be applied, and from lack of an exhaustive list of spectral indicators that may be used for uniqueness tests. (2) As spectral indicators we consider: emission lines, self-reversed emission cores, the presence of He I lines in stars not of early type, coronal-type high ionization, excess continuum emission in the rocket UV and the far infrared. (3) To determine what we can infer from observations, we summarize information: inferred by comparison of models to data, on velocity fields, and on spectral variability which might suggest chromospheric activity. (4) We summarize the evidence for chromospheres in A stars, as being those where convection-induced acoustic heating is marginal.