From  John E. Keller's remarks, while not in full disagreement  Llorens Cisteró , one senses the implication that the practice was from time to time and that "even now  [cantigas  are] sung [at the cathedral of Seville] to musical accompaniment."^[4] Thus, any attempt to verify whether or not Alfonso's command was executed per annum, through the centuries, would necessitate a laborious and time-consuming perusal of the cathedral's  Actas capitulares,  particularly the entries for the feast days of the Virgin: February 2, March 25, August 15, September 8, October 12, December 8. ^[5] Apart from Seville and Toledo where  Alfonso  held his court, information concerning when, where else, and how Alfonso's  Cantigas were performed during his reign has to date yielded very little  descript ive documentation of musicological significance, except for scanty iconographical evidence.

One has only to view the physical appearance of any one of the extant [160] Cantigas  codices bearing musical notation-ranging from 315 to 485 mm long and from 217 to 326 mm wide, and containing from 160 to 370 leaves of parchment-to realize how unwieldy each was for the solitary singer as well as the clarity of its notation if additional singers were huddled before it. This immediately calls to mind the larger and cumbersome cantorales, or choir books, such as those resting on facistolia (lecterns or choristers' desks) in the choir loft of the royal monastery of  El Escorial,  around which several singers stood to sing from the same musical notation, inscribed on huge sheets of parchment. Whether such a group gathered around a Cantigas  codex, or whether one singer, knowledgeable in reading its notation, taught selected texts and tunes to those gathered around him, remains an enigma. One can add to this problem such further speculations as those pertaining to performance practices, including dancing, with or without the accompaniment of musical instruments like those depicted in several of its miniatures, ^[6] as well as the manner of accompaniment.
 

A closer study of the text-tune relationships of both the Marian miracles and the praises sung in her honor may even enable us to differentiate the poetic texts, for which their tunes were created simultaneously, from texts which may have been suited to known melodies of the time taken from either liturgical or secular sources. At the same time, we can take into account their varied formal, modal,  cadential , and intervallic structures,ambitus , syllable count, rhythmic features, mensural schemes, and versification. ^[7] The melodic origins of the  Cantigas tunes await further investigation.^[8] Moreover, considering the fact that the Galician-Portuguese dialect was neither the spoken nor even the written language of most Castilians, it is unlikely that even a handful of sung  cantigas  would have survived in oral tradition.
 

Cantiga 10: "Rosa das rosas"
 

Still, the possible instance of a cantiga tune surviving in oral tradition up to the third decade of the present century appears to have given rise to a few casual notices, even though its sung text is a Castilian translation of the original Galician-Portuguese. I am referring, of course, to  Alfonso  X's cantiga de loor  10 , "Rosa das rosas" (Figure 11-1),  which appears as no. 263 among the musical notations in Kurt Schindler's Folk Music and Poetry of Spain and Portugal, published posthumously by the Hispanic Institute of Columbia University in 1941 ^[9] (Example 11-1b). Schindler collected the [161] cantiga  in the town of  Ceclavin, in western  Cáceres  near the Portuguese border, during his short visit there in the late summer of 1932. This was the period of Schindler's second field trip to Spain, ^[10] during which he transported a Fairchild portable recording apparatus for the purpose of gathering his material on aluminum discs. The label on disc no. 150A, containing "Rosa [de  las r]osas , " cites  Amado Vives Amores  as Schindler's informant. Furthermore, in the text portion of his published collection, Schindler identified item 263 as  "la Cantiga  X del  rey  Don  Alfonso el sabio, "  yet he questioned whether it was  "tradicional en Ceclavín." ^[11]

Daniel Devoto,  in his critical review, was the first to take note of its inclusion in Schindler's magnificent collection. He concluded that the text is traditional (folklórico)  and relates to its original antecedent, the  cantiga, despite the difficulty of proving it  (es [de no establecerse una transmisión no tradicional ] folklórico, y cuenta con un antecedente, la cantiga original) . ^[12] Citing  Devoto's reprinted version, Jacques  Chailley concurred that "the cantiga 'Rosa das rosas'  is preserved in Spanish oral tradition, and in Kurt Schindler's [book] it was collected as a popular song" ( s'est conservé dans la tradition órale  espagnole,  et a été recueilli comme chant  populaire dans Kurt  Schindler). ^[13] I initially agreed, including in my article not only Schindler's but also four additional modern transcriptions of  "Rosa das rosas"  for comparative purposes. ^[14] In a recent critical study of Martin Codax's Cantigas de amigo,  Manuel Pedro  Ferreira  included (in the first of three appendixes) a short discussion of  "Rosa das rosas, "  wherein he too expressed a similar view, while, at the same time, suggesting that the  cantiga  may have had its own manuscript tradition. ^[15] In support of this last position, Ferreira  cited  Luis  Villalba's published arrangement of  "Rosa das rosas"  as the "manuscript" source of the Schindler tune, basing his evidence not on the tune, but on the comparison of the text  underlays of their initial strophes. ^[16] The text for  Villalba's setting had been  Castilianized by R.  de Valle;  the deviations in Schindler's text are enclosed in brackets:

Rosa en que el Señor puso su querer

Flor la más hermosa que se vió nacer

Virgen que hace dulce todo padecer

Amor que hace nuestros sus santos amores.

---

---

[165] Had  Villalba known of the existence of the earliest Castilian version of the text, recently discovered by John Keller, ^[17] he might have placed it below the original text of his transcription as he had done with R. de  Valle's translation. Thus, if both Schindler's and  Villalba's text  underlays had agreed closely with the medieval Castilian version, it would have indeed added weight to the speculative opinions registered above for an unbroken chain of oral transmission.

According to Keller, the Castilianized poetic rendition of the original can be considered one of the earliest examples of thirteenth-century Castilian poetry, the first three stanzas of which are almost literal translations from the Galician: