196 JOURNAL OF THE ARNOLD ARBORETUM [vol. iv 



utrinque laxe reticulata, 1.5-2 cm. longa et 7-9 mm. lata vel ad 2.5 cm. 

 longa et ad 1.2 cm. lata; folia ramulos novellos fulcientia ultra 3 cm. longa, 

 ovato- vel obovato-oblonga, in petiolum longiorem attenuata, surculorum 

 elliptica vel ovato-elliptica, basi cuneata breviter petiolata, ad 4.5 cm. 

 longa et ad 2.3 cm. lata, grossius serrulata, subtus glaucescentia. In- 

 florescentiae breviter racemosae vel fasciculate- (vel subumbellato-) 

 racemosae, rarius elongatae et nutantes, 1-3 cm. longae; flores aperti 

 nondum visi. Fructus elliptic!, coccinei, circ. 8-9 : 5 mm. magni, fere 

 semper esty lares. 



According to Schrader, whose type I have seen, he received this form from 

 Audibert at Tarascon and also as B. ilicifolia from the Botanical Garden 

 at Vienna. I have seen a "catalogue des Pepinieres des Freres Audibert 

 a Tonelle, pres de Tarascon" of the year 1831/2. Here a Berberis vulgaris 

 v. provincialis "de provence" is mentioned. This seems to point to a 

 wild form of B. vulgaris from the southwestern parts of France, but I 

 have not found in books any mention of a wild form agreeing with Schra- 

 der's description. In a note he says : " B. provinc. forte ad prim. sect, refer, 

 cum ramis minus rubris et flor. cum vulgaris I am, however, not sure 

 whether the form from Vienna and that from Tarascon really were of the 

 same origin. As I stated in 1906 the form preserved in Schrader's Her- 

 barium is very similar to the form described by Koehne as B. serrata, 

 and there is a reasonable chance that both are of the same origin. The 

 true B. serrata is a very distinct form recognized at the first glance from 

 B. provincialis by its very acute leaves. In the Arnold Arboretum, 

 however, I found, under no. 5886-1, plants raised from typical B. serrata 

 (no. 18 ex Rochester) that show a closer similarity to B. provincialis than 

 to typical B. serrata (as represented by Herb. Dendr. Koehne no. 463). 

 I therefore feel justified in regarding B. serrata as a mere variety of what 

 I consider according to Schrader B. provincialis. It would be interesting 

 to raise from seeds several generations of this hybrid in order to trace its 

 real parents. One of them is doubtless B. vulgaris. The presence some- 

 times of three ovules, the character of the inflorescence, the shape of the 

 petals, the appearance of the leaves and spines and the color of the twigs 

 indicate B. sibirica as the second parent, but the typical B. emarginata 

 has larger flowers and inflorescences. After all, B. provincialis may be of 

 very similar origin as B. emarginata. 



Berberis provincialis var. serrata, nov. var. 



B. serrata Koehne, Deutsche Dendr. 170 (1893). 



B. microphylla var. serrata Hort. Kew, in a MSS-list of 1884. 



Frutex ex auctore 0.6 m. altus, ramis erectis squarrosis; ramuli annotini 

 flavo-rubri vel satis rubescentes, distincte sulcati, laeves, etiam biennes 

 cinereo-brunnei, vetustiores nigrescentes; internodia 1-1.5 cm. longa; 

 spinae graciles, 1-3-fidae, flavescentes, vel rubescentes, mediae ad 10 mm. 

 longae, subtus sulcatae. Folia 4-8-fasciculata, inaequalia, matura chartacea, 



