Epitype can not be a part of original material. You can see my paper - 'Epitypification of Goodyera affinis (Orchidaceae)' published in The Journal of Japanese Botany, vol 88 (available in my ResearchGate profile) which can help you in this context. Please fell free if you have any doubt. 

An epitype can be designated only when available "regular" type material (an epitype cannot exist alone!) is not adequate to correctly apply a name without ambiguities.

Now, if you face this problem during your own study aimed to designate a type, and some of the original material is more complete, then this is the best choice: you can directly select that specimen as the type (in this case your problem simply does not exist). The words of ICN "or all original material" refers actually to a situation like this: if you have all the original elements which are incomplete, you are anyway forced to select one of them as the lectoype! In that case, an epitype is very advisable to avoid ambiguities.

If, as I argue may be your case, someone validly designated a lectotype for that name BEFORE your study, well, unfortunately this is the definitive type (if correctly selected among original material).

If subsequent studies show that this lectotype does not allow an unambiguous application of the name, well an epitype is worth to be selected in support of that lectotype. But, at this point, it makes no sense to select another element of the original material! The best thing, in that case, is to select a well preserved, complete, recent collection, from a well defined geographical area, which according to the epitypifying author can definitely help to apply that name without ambiguities.

If you have further doubts, please feel free to contact me!

Thank you very much Dr. Peruzzi for your clear answer, which explains the rationale in the Code very well. Certainly, if lectotypifing authors always had a precise knowledge of all of the original material, as the ICN assumes, some regrettable choices I have seen would have been avoided. Fortunately my question was purely theoretical, it doesn't refer to any real world case I encountered. I hope we will have the chance to meet in person.

Regarding to the last paragraph of Dr. Peruzzi´s response: " it makes no sense to select another element of the original material! The best thing, in that case, is to select a well preserved, complete, recent collection,"

If a previous selection of lectotype  is demonstrably ambiguous and cannot be critically identified for purposes of the precise application of the name to a taxon...

For any reason a lectotype/holotype might have been damaged or it might have been just a bad selection...

Why a recent collection would represent better the name than, for example, a very complete and well preserved specimen amongst the avilable syntypes (If them existed)?. I don´t see any restriction in the code about it. 

Regards

A bad selection is however definitive, with the exception of those cases falling into proposals to "conserve types" under the Code.

If an epitype is needed, the better is its quality and definition, the higher will be its potential in helping a correct application of the name. Original materials (especially those from pre-XX century) may lack informative portions or other useful details. For this reason I stated that "it makes no sense" (= it is absolutely NOT necessary!) to select an epitype from the remaining original material. Of course, the are no restrictions from the Code in this sense.

Ok Lorenzo, 

Then I guess, the answer to the original question is yes. An epitype can come from original material. 

Of course, each case may need a different approach and a different response. Probably, in many cases, selecting more recent and complete material will be the best option. However, for some names, it might be better to choose an epitype amongst the remaining syntypes. Always in the case they are complete and well preserved. 

Regards 

Yes Jose, the answer to the original answer is "yes". But I argue that the real sense of Moreno's doubts were actually about the exact meaning of the sentence "or all original material" within Art. 9.8 of the ICN.

However, the things can be even more complicated! Recently I typified the name Asperula crassifolia L. (Taxon 62(4): 827-829, 2013), whose lectotype (a forced choice) does not allow a clear-cut, unambiguous application of the name (which is moreover described by Linnaeus for a wrong place "Habitat in Creta"!). Hence, we selected as epitype for that name a complete and representative (but ancient) specimen which at the same time is also the lectotype for the name Asperula tomentosa Ten. (a later heterotypic synonym, described from Capri island - S Italy - the only area where this species is known to occur at present)!

Lorenzo is right, my question was precisely on the meaning of "or all original material", and I will get back to it. While I reckon I now understand it, I still think that there is a problem there. Let me reword what I understand is our collective conclusion:

that sentence doesn't really apply to the hypothetical case I was proposing. In fact, it doesn't refer to the process of epitypification at all: it only serves to remind that, in case no lectotype has already been selected, before bothering with epitypes, one must try and find an unambiguous specimen from the original material and designate it as a lectotype. Which is already explained in detail throughout article 9. Only if the lectotype or potential lectotypes are ambiguous, an epitype should be chosen, from whatever material is best suited to solve the problem. Whether the epitype should come from original or new material, it is for typifying author to decide (and we all agree that, most of the times, it will be new).

If this interpretation is correct, the sentence "or all original material" is superfluous, since the fact that a lectotype should be chosen before or at least together with an epitype is made clear enough in the very same sentence. It is also potentially confusing, since it could easily lead to the wrong conclusion that no original material can ever be chosen as an epitype, for some inexplicable reason, whether a lectotype is still to be chosen or it was chosen already. This was said to me more than once.

On the other hand, in the unlikely case that our interpretation is wrong, then the true meaning of Art. 9.8 has eluded our collective wits, which would mean that it is so obscure that it must be urgently amended.

Or I just got something wrong (not unlikely). Please, let me know what you think!

Hi Moreno, 

I don´t really think the sentence "or all original material" is so problematic. The key is the "or", separating each of the cases: 

1.- "when the holotype, lectotype...is demonstrably ambiguous and cannot be critically identified for purposes...". 

In this first case an epitype could be chosen from any other existing syntype, or from new material. As I said before, I think not always new material will represent the name better than a complete and well preserved syntype. 

2.-"or previously designated neotype...is demonstrably ambiguous and cannot be critically identified for purposes...". 

In this case, no "original material" exists acording 9.7, so only new material could be used as epitype. 

3.- "or all original material associated with a validly published name, is demonstrably ambiguous and cannot be critically identified for purposes..."

This last case seems to be the problematic one. I see mainly two possible cases:

- There exists an holotype or a lectotype (in need of an epitype). All other original materials "isolectotypes, isotypes, syntypes" lack the necessary features to be selected as epitypes. Then only new material can be choosen. 

- There exists no holotype. If a lectotype is selected, it must be selected amongst the original material, even if it is incomplete, and then select an epitype from different materials. 

Anyway, this is being a nice topic to discuss. 

Regards

I think I don't really understand your table, I'm sorry. This is what I understand are the cases discussed in Art. 9.8:

You should select an epitype in these cases:

1) There is a holotype, and it is ambiguous

2) There is a previously designated lectotype, and it is ambiguous

3) There is a previously designated neotype, and it is ambiguous

4) There is no designated type, and all of the original material is ambiguous.

In case 4, you should first select a lectotype from the original material and then an epitype from wherever you want. This may be done in a single publication, of course, but conceptually it is a two-step procedure. In other words, I was arguing that the sentence is superfluous since, if no type already exists (= cases 1, 2, 3 don't apply), you first have to refer to previous points in the same Art. 9 and make them apply.

I argued also that point 4 is confusing, with that wording, because it can be read as:

4) An epitype can be chosen only if all original material is ambiguous (and therefore out of that), regardless of whether a designated type already exists or not.

Of course, if you think about it, this wrong interpretation creates a "hole". If it were true, you could have a name that is supported by some unambiguous original material which cannot have type status (= neither lectotype nor epitype), which is is plain contrast with the general working of the Code. But this is not immediately obvious to understand at all.

It is a nice topic indeed, but maybe a bit pointless!

Moreno

PS. Incidentally, I think zoologists allow such "holes" and give them a specific type status: that of paralectotype.

Sorry Moreno, but I think your interpretation of the sentence "or all original material" is wrong, and therefore leads you to wrong conclusions: 

Let´s treat it as what you called case 4:

As I wrote in previous responses, from my point of view "all original material" does not exclude explicitly the existence of an holotype or a lectotype. It just implies that (if exist) they are ambiguous, as well as any other original materials. 

Of course "all original material" also would apply, as you say, to many cases where no holotype or lectotype exist. In this case, as it can be deduced from the final sentence of Art. 9.8 (Designation of an epitype is not effected unless the holotype, lectotype, or neotype that the epitype supports is explicitly cited), you will need first to select a lectotype from the original material, and then an suitable epitype. 

I see no problem in taking two steps if needed. However it does not mean that you "make cases 1 and 2 apply". In cases 1 and 2, other original material would be suitable to be chosen as an epitype, and therefore they are different to case 4. 

Regards, 

Jose

I thought about your interpretation very carefully. It definitely makes sense, covers all possible cases and leads to the correct conclusions, while finding a good explanantion for "or all original material". I'm still not really convinced, though, that that's how Art. 9.8 is meant to be read, it seems to me too convoluted, but I might very well be wrong. How I read it leads exactly to the same conclusions, but makes "or all original material" somewhat redundant (but, thinking it over, less so than I originally thought). In any case, since our conclusions are the same, I think arguing any further between ourselves will not be useful. Thank you very much for your input.

Moreno