( 135 ) 
eeiyed the excellence of the Linnean method. 
Mr. Curtis has, in many genera of this diffi- 
cult clafs, difcovered conftant marks by which 
they may be diftinguifhed in different ftates 
of growth. In the onopordon acanthium, 
when the flowering is over, he has obferved 
that the innermoft fcales of the calyx clofe 
ftrongly together, and preferve the feed, con- 
trary to the calyx of carduus, and moft other 
genera of the compound flowers, w r hich, as 
has been before remarked, expand and difperfe 
their feeds. The fmaller flowers of this clafs 
are more difficult to inveftigate ; but, if pro~ 
ceeded with in the fame manner as the larger 
kinds, a competent knowledge of them may 
foon be obtained. A numerous tribe of plants, 
termed the urnbelled plants, which are con- 
tained under the clafs Pentandria, will be 
found more eafy of accefs to the young bo- 
tanift if he has fome previous information in 
the mode of their investigation. The um- 
belliferous plants fhould be gathered for ex- 
amination before their florets are wholly 
expanded, otherwife it will not be eafy to de<° 
termine the clafs to which they belongs as the 
anthers frequently drop, off as foon as they 
arrive at maturity. If this is attended to, it 
& 4 will 
