332 The Botanical Gazette, [September 
must be regarded as departing most widely from the primitive 
type. The simple pistil, developed from a single phyllome 
union be with one another or with other structures, must be 
derived and consequently higher. So too when all parts of 
the flower are separate it is a primitive condition, and when 
they are united it is a derived structure. ; 
Applying these principles to the flowering plants ithe 
comes evident that in the dicotyledons either the Apetale ot 
the Polypetalz must furnish our starting point. The Game 
parts, and the entire omission of others. _ It will not be die 
to determine that the Ranales must take rank below all 0 
Polypetalz, in the sense of representing more nearly 
any other group the primitive dicotyledons. a 
The attempt to make a natural system by linking 
family in a long undilating chain, by concatenatio®, wee 
scientific because it absolutely fails to conform to oe 
al. 
us not. cling to the old because it is, inconv 
let us not cling to it through a mistaken ‘aeaee 
practice of the fathers, let us not cling to 1t as — 
may be found in a new system. Science 1s | 
the old, when the old is no longer the true; nts the ® 
work of years, when that work no longer F Pre system 
and it dares to reach out and frame a rationa hype neti 
though some parts of it for atime rest Upon UF 
grounds. :.  Piaieale a 
A revised arrangement of the Benthamian ‘‘series 
' ering plants. 
Monocotyledons. : 
I. Apocarpe. 5. Glumace®. jrock 
2. Coronariez. 6. shag (Hy 
3. Nudiflorze. 7. Epigya 7 
4. Calycine. 8. Microsperm® 
