120 MB. a. BE^'THAM ON MTRTACE^. 



Las also been well brought forward by Brougniart and Gris in 

 their papers on New Caledonian Mj^rtacesej although it may be 

 doubted whether it may not have led them rather too far in the 

 dii>tinction of genera allied to Metrosideros, 



5. Fiturr. 



The fruit of Myrtacea? gives very important general but not 

 absolute characters, dividing well the Avhole order into four great 

 types: — the small one-seeded indehiscent nut of Chamselauciese ; 

 the capsule opening loculicidally at the top of Leptospermese ; 

 the indehiscent bcxTy, or rarely drupe, of Myrtese ; and the hard 



indehiscent or operculate, usually large or several-seeded fruit of 



Lecythidese ; but in each of these tribes there are exceptions 

 chiefly monotyj^ic or very small genera, or isolated species. And 

 beyond this general distinction it is very rarely that modifications 

 of the fruit are at all available as Reneric characters. 



6. Seed. f^ 



The seed, apart from the embryo, is also of very little avail. 

 Its insertion, dependent on that of the ovule, is, of course, Mhen 

 undisturbed, equally important ; but it is so frequently modified 

 by the growth of the fruit, that it is always much safer to ascer- 

 tain it at the time of flowering. The size and shape of the seed, 

 tlie texture of the testa, its expansion into wings of various 

 shapes, the presence of a small quantity of albumen or its total 

 absence, are very rarely, as far as hitherto know^n, constant m 

 large genera, and will probably be found still less so when more 

 species shall have been examined. 



We now come to the great character of the embryo, to the 

 modifications of which the greatest importance has always been 

 attached, owing, in some measure, to a priori reasoning, the 

 danger of which is fully exemplified in Myrtacese, To this day I 

 do not believe that the embryo has been seen in one-half of the 

 published species of the order. There are considerable genera, of 

 Avhich I have had hundreds of specimens before me in various 

 sta^^^es of flower or fruit, of whicli the embryo had never been 

 described, and in which I have only been able to see it in tw o or 

 three seeds, or in which it is still unknown; and the fallacy of 

 presuming upon the embryo from analogy in other respects is 

 shown by tlie numerous species which the most acute systematic 

 botanists had placed in the wrong genus before the embryo was 

 ascertained. In the tribe Myrtese, indeed, if we get the ripe 



