Commentary Par XV 22-24

That is, in its approach before its descent, this 'star' follows the right angle made by the 'arm' and the 'stem' of the cross.  For Dante's earlier treatment of the cross of Mars, see Conv.II.xiii.22: 'This is also why in Florence, at the beginning of its ruin, there was seen in the sky in the shape of a cross a great quantity of these vapors which accompany the star of Mars' (tr. R. Lansing).

The apsidal chapel dedicated to the local martyr of Ravenna, St. Apollinaris, in the church dedicated to his memory, S. Apollinare in Classe, displays over the altar a mosaic cross that has jewels depicted as being embedded in it, with Christ's face as the central jewel, located at the transverse of its two elements.  Is it possible that Dante was thinking of that particular cross?  We do not know whether he visited Ravenna before he moved there, ca. 1317-19, nor when he composed these cantos (but see the last paragraph of C.Par.IX.29-30, reporting on the possibility that that canto reflected the pressure of the recent past, the years 1314-1315).  Petrocchi's dating of the composition of the last cantica would have it begun ca. 1317.  And so it is at least possible that the greater portion of Paradiso was composed in Ravenna.  At any rate, Jeffrey Schnapp (Schn.1986.1), pp. 171-203, offers detailed arguments that locate various architectural features of Ravenna in the texts of the central cantos of Paradiso, including the mosaic found in this chapel, with its jeweled cross (pp. 180-85).  And see Desideri (Desi.2001.1) for another effort to document the influence of Veronese artifacts on Dante.