THE CHROMOSOKIE NUMBERS OF THE BRITISH MENTHAE 247 



v.c. 17; 2n = 84 from v.cs. 17, 65, 66 (4) and 67; 2n = 96 from v.cs. 66 and 67 (2); 

 2n 108 from v.cs. 67 and 104; 2n = 120 from v.cs. 40, 67, 70, and 77. The plant 

 with a diploid number of 54 was referable to M. X gracilis and the one with a diploid 

 number of 60 was referable to M. X cardiaca. Further material of both these is 

 required in order to determine whether these numbers are constant. The diagram 

 below indicates the main hybrid combinations and recombinations which ajre theoretically 

 possible between M. arvensis and both cytotypes of M. spicata. It is noteworthy that 

 both forms of M. spicata can take part in the formation of the M. X gentilis complex, 

 though it appears that the one with 2n = 48 is most frequently involved. M. X gentilis 

 with 2n = 84 is known to set fertile seed on occasions, and young plants have been 

 raised from such seed, but these had not flowered up to the time of writing. Hybrid 

 swarms are frequently encountered and the ones examined have been combinations of 

 M. X gentilis (2n = 84) and M. arvensis. The plants in such swarms show every inter- 

 mediate between M. X gentilis and M. arvensis and frequently set well-formed seed. 



M. arvensis 

 -72 



M. spicata 

 36- 



M. arvensis 

 -72 



66 / 

 2n 



96 

 gentilis 



M. spicata 

 48- 



60 

 cardiaca 



x2 



120 

 gentilis 



/ 



54 



In 



84 

 gentilis 



M. X muelleriana F. W. Schultz. Material of this very rare mint, from its original 

 Devon locality, was found to have a diploid chromosome number of 60. M. arvensis 

 and M. rotundifolia are obviously the parents involved in this hybrid, and the above 

 chromosome number 'indicates a back-cross of the Fj hybrid with the M. arvensis parent. 



M. arvensis M. rotundifolia 



72 28 



muelleriana 



M. X nemorosa Willd. Morphological characters suggest a M. longifolia X M. 

 rotundifolia parentage for this mint. Plants belonging to this hybrid frequently so closely 

 resemble the former parent that so far I have been unable to separate them. A diploid 

 number of 36 was obtained in material from v.c. 91 which is definitely referable to this 

 hybrid. The presence of at least two cytotypes (2n = 36 and 48) in M. longifolia, one 

 of which (2n = 36) is the same as would be expected to occur in M. X nemorosa, prevents 

 the separation of the species from its hybrids on a cytological basis. As already mentioned 



