>1.) Is the holotype a single specimen or can it be more than one for
>plants?
The holotype is one single element (i.e. mostly one individual or several
fragments of one individual, in some cases several individuals which are
obviously clones of each other). It is rather dangerous (but not forbidden)
to consider all specimens on one sheet to belong to the same plant (or
clump of plants or isolated population) and to designate the whole sheet as
a (holo- or iso-) type specimen (you will probably see a fine isotype of
_Drosera uniflora_ at P which consists of quite a lot of individuals).
>2.) When you publish a new plant species, should the description be
>based on a single holotype or on various specimens which the author
>considers to be conspecific? ......2 descriptions?
As long as only a single element is unambiguously designated as the
holotype, the description may be based on whatever material the author
chooses to belong to the new taxon. It would be fine if the author cited
the other (i.e. non-holotype) specimens investigated.
>3.) In the case of various holotype specimens, would the species be
>invalidated if in the future someone decides that those specimens
>represent 2 or more distinct taxa?
Not normally. The element from the original (mixed) material which
corresponds most to the distinguishing features mentioned in the protologue
(or illustrations published together with the protologue) has to be
selected as the lectotype of the name originally based on mixed specimens,
which is thereby emended (and stays valid). The remaining material may be
assigned to another (and possibly even a new) taxon.
Only in very few cases in which the ambiguous name has already caused
severe and persistent confusion (so restitution of the ambiguous name would
cause an unfavourable nomenclatural change), it may be proposed for
rejection so that unambiguous names can take the place even if they both
were published later.
>4.) In the case of a single holotype specimen, would you have to
>write 2 descriptions in your paper, one based on the holotype and
>another based on paratypes and syntypes (the latter for additional
>info and also to give the details which can not be measured from the
>single holotype which can not be taken apart)?
This is not necessary. You may (e.g. in cases where measurements or
features mentioned in the protologue do deviate from the data obtained from
the holotype) indicate which specimens the data were obtained from.
Kind regards
Jan