301 



being examined IVoni Ihe laller alone. Iii addition, dilTercnt 

 species from Ihe bolanical gardens of C.openhagen and Chrisliania, 

 sonie sponlaneous swedish bushes, and some forms from private 

 gardens, aniong thein the garden of Ihe eminent rhodologisl Rev. 

 R. Matsson, Hudiksvall, Sweden, have been examined. This makes 

 a total of about 300 bushes representing approximatelj' 230 dilTe- 

 rent forms. The determination of my material — I have pressed 

 shoots of all lixed forms — bas been controled by the distinguished 

 swedish expert in roses S. Almquist. For most of the species, I 

 have conlined the investigation to the development of the pollen; 

 for many apomictical forms, I have as well invesligated the reduc- 

 tion division in the ovules. As I hope soon to have an opportn- 

 nity of presenting the complete report of my studies, I shall here 

 l)rielly sum up the main results and give some of my conclusions 

 regarding the chief /tosa-problems. 



1. The fundamental haploid chromosome number in Rosa is 7, 

 not 8 as was stated by Strasburger. 



2. The genus Rosa may in cytological respect be divided into 

 two very dilTerent groups of species: on one side the very poly- 

 morphous Cani/ja-section (in the widest sense) including such sub- 

 seclions as Caninae (canina + diimeforum), Afzelianae [glauca + corii- 

 folia), Villosae, Tomentosae, Riihiginosae, Agrestes, Junzillianae, Slijlo- 

 sae and Riibrifoliae: on the other side all the other sections. The 

 latter group is distinguished by the occurrence of only paired chro- 

 mosomes in the diakinesis: 7, 14, or 21 pair. The former, the 

 Ca/j/na-section, which contains the greatest part of the 7?osa-vege- 

 tation of Europé, is characterized by the peculiarity of showing in 

 the heterotype division both bivalent and single chromosomes, 

 usually in the proportions 7 paired plus 14 unpaired, or 7 paired 

 plus 21 unpaired, or 7 paired plus 28 unpaired, thus showing an 

 analogous chromosome condition to what was stated l)y Rosknheiui 

 in the well-known Drosera-baslard and låter has been found in some 

 other hybrids. 



3. Considering only the somatic number of chromosomes, I 

 have in my material found the following numbers: 14, 21, 28, 3ö 

 and 42. Further, in some specimens, I met with numbers not 

 being a multiple of 7 (anorthoploid forms). These bushes may be 

 explained as hybrids representing a låter generation than F^ (see 

 the points 20, 21 and 25). 



