Can you provide an author/submissions guide for one such journal so we can see exactly what this should look like? Also post some complete references of this sort here as well.

That does help.

So we need to do something like

[1] a) B. M. Trost, Chem. Eur. J. 1998, 4, 2405–2412; b) H. J. Ache, Angew. Chem. 1989, 101, 1–21; Angew. Chem. Int. Ed. Engl. 1989, 28, 1–20; c) H. Frey, Angew. Chem. 1998, 110, 2313–2318; Angew. Chem. Int. Ed. 1998, 37, 2193–2197.

I'm afraid that this isn't yet possible, but hopefully one of our CSL gurus can confirm that. If this is indeed not currently possible, it probably won't happen for some time, since I don't see mention of such behavior in the CSL 1.0 specification, and I don't think that CSL 1.1 has even been started.

But maybe another of the style requirements, standard journal abbreviations, will also make it into the CSL 1.1 standard, so at least those two birds will be hit by a single stone, even if that stone is relatively far off.

If this isn't possible now, we'll want as large as possible a listing of journals that require this and similar behavior, to justify and correctly develop such a feature for a future release. Can you survey journals in your field to create such a list?

Yes, you understand right.
A single reference will also get a number (like a group) but in the bibliography it won't have a letter index. like this:

[4] a) J. S. Lee, P. A. Ulmann, M. S. Han, C. A. Mirkin, Nano Lett. 2008, 8,
529–533; b) Z. S. Wu, S. B. Zhang, M. M. Guo, C. R. Chen,
G. L. Shen, R. Q. Yu, Anal. Chim. Acta 2007, 584, 122–128;
c) F. X. Zhang, L. Han, L. B. Israel, J. G. Daras, M. M. Maye,
N. K. Ly, C. J. Zhong, Analyst 2002, 127, 462–465.
[5] S. Chah, M. R. Hammond, R. N. Zare, Chem. Biol. 2005, 12, 323–
328.

I will search for more journals that prefer this kind of citing and supply a list ...

I can confirm that this is, unfortunately, quite common. It has come up a bunch of times as I have coded styles in the sciences and engineering. While I agree it's a terrible idea, it probably is something that csl should be able to do.

Yes, I think so.

And they all do them in the same way? Are references sorted within the compound reference? Is ibid or some other form of shortening used?

But the list confirms that organic chemists and chemists in general, at least in Europe, like this style.

From the hechtlab.de PDF preprint provided above:
Text:
Furthermore, the 1:2 stoi-
chiometry of the formed complex was verified by a Job
plot[41] (Figure 16 in the Supporting Information), and the
shape of the actual Job curve, which is known to depend on
the complex stability in a characteristic manner,[41c] was in
excellent agreement with the stability constant of 108 mÀ2 de-
rived from the ITC measurements.

Endnote:
[41]a) P. Job, Ann. Chim. 1928, 9, 113 – 203; b) C. Y. Huang, Methods
Enzymol. 1982, 87, 509 – 525; c) L. Sommer, M. Langovμ, Crit. Rev.
Anal. Chem. 1988, 19, 225 – 269.

And the last sentence:
This should lead to improved properties in supramolecular
polymeric materials,[1b, 45] as the self-healing is accelerated.

Endnotes:
[1] Original work: a) G. T. Morgan, F. H. Burstall, J. Chem. Soc. 1932,
20 – 30. A comprehensive treatment is given in: b) U. S. Schubert, H.
Hofmeier, G. R. Newkome in Modern Terpyridine Chemistry, Wiley-
VCH, Weinheim, 2006.
..
[45] a) T. Nakano, Y. Okamoto, Chem. Rev. 2001, 101, 4013 – 4038;
b) J. J. L. M. Cornelissen, A. E. Rowan, R. J. M. Nolte, N. A. J. M.
Sommerdijk, Chem. Rev. 2001, 101, 4039 – 4070.

Thus, it appears that we have backreferencing to subitems, endnotes, and no separate bibliography.

So am I understanding that the details of this case are so complicated that, if implemented, it would force the processor to create fake endnotes just to be able to represent it?

If yes, then this would suggest it's too complicated to bother. I seriously doubt that "Endnote is able to do [all of] this since 2009," as jens suggested in the OP.

@adamsmith--
No, we need to use things like [1b,45] in-text. As Frank points out, word processors don't let us create customized indices for endnotes, or at least the current plugins can't.

that's what I meant, I think. Aren't all numeric styles (by which I mean Nature, NLM, IEEE etc.) just in-text styles? There doesn't seem to be something extraordinary about the citation - I'm much more concerned about the bibliography.

A chance, yes, but it's not going to be soon.

I am a chemist and require the use of composite references for every major chemistry journal. Here is a list of examples, just off the top of my head:

- Journal of the American Chemical Society
- Organic Letters
- Journal of Organic Chemistry
- Chemical Reviews
- ACS Catalysis
- Chemical Communications
- Green Chemistry
- Org. Biomol. Chem.
- Chemical Science
- ChemBioChem
- Angewandte Chemie
- Angewandte Chemie International Edition
- Advanced Synthesis & Catalysis
- Chem. Soc. Rev.
- ChemCatChem
- Chemistry - An Asian Journal
- Canadian Journal of Chemistry
- Chemical Letters
- Australian Journal of Chemistry
- RSC Advances
- Synthesis
- Coordination Chemistry Reviews
- Dalton Transactions
- Synlett
- Org. Proc. Res. Dev.
- Journal of Molecular Catalysis: A
- Acta Crystalographica
- Tetrahedron
- Tetrahedron Letters
- Tetrahedron: Asymmetry
- Organometallics
- J. Med. Chem. Lett.
- ChemMedChem
- ChemPhysChem
- New Journal of Chemistry
- Natural Product Reports
- Organic Chemistry Frontiers
- Inorganic Chemistry Frontiers
- Catalysis Science & Technology
- ChemSusChem
- Journal of Organometallic Chemistry
- European Journal of Organic Chemistry
- Inorganic Chemistry
- Journal of Physical Chemistry A
- Journal of Physical Chemistry B
- Journal of Physical Chemistry C
- European Journal of Inorganic Chemistry
- Asian Journal of Organic Chemistry
- European Journal of Inorganic Chemistry
- Journal of Medicinal Chemistry
- Langmuir
- Molecules
- Beilstein Journal of Organic Chemistry
- Arkivoc
- Journal of Fluorine Chemistry

... and many, many others.

I promise you, I came up with that list from memory - they all use composite references. These are not obscure publications, either. They cover the spine of the literature in organic and inorganic chemistry (and there are probably many others from physical and materials chemistry which I didn't mention). Composite references are used across the major publishers also. It's not a European thing, as someone suggested above, it's international. It's used by the ACS, RSC, Wiley, Springer, Thieme, Elsevier and other publishers.

EndNote can handle composite references, which testifies to the importance of this way of doing things.

There is probably some difference in how the value of composite references is perceived between people from different backgrounds. However, in chemistry it is enormously logical to use it because it allows many references to be grouped by topic, for example in writing succinct introductions, or referring to multiple works by a single author/research group.

This feature is the only thing that stands between me and a move from EndNote to Zotero, and it has done for over five years. It's probably the same thing for innumerable other chemists (I suspect quite a few thousand), who would prefer to use Zotero's otherwise fantastic feature set over EndNote's expensive version.

Please, please, please develop this feature.

Please.

THE LOGIC OF COMPOSITE REFS

So "rules" are fairly simple for these types of references. I found the above discussion about this overly complex. Here is how it works.

If you are writing with composite references enabled (e.g. in EndNote), then this happens:

Imagine you are writing an introduction three sentences long. Something that would ordinarily (for non-chemists) look like:

----

"The construction of C-C bonds is an important area of chemistry.[1-3] The Suzuki-Miyaura reaction ranks amongst the most important catalytic processes for this purpose.[4] To date, various mechanistic rationales for this reaction has been proposed.[5-6]"

References

1.
2.
3.
4.
5.
6.

----

Note that references 1,2 and 3 as well as 5,6 are inserted grouped together, hence the ranges but these appear separate in the Bibliography (or References) section. When in Endnote I insert a reference and then another directly afterwards (or select two to insert at the same time), they are grouped.

For a chemist writing with the composite references option, the reference manager would make it look like this:

----

"The construction of C-C bonds is an important area of chemistry.[1] The Suzuki-Miyaura reaction ranks amongst the most important catalytic processes for this purpose.[2] To date, various mechanistic rationales for this reaction has been proposed.[3]"

References

1. a)..., b)..., c)...
2.
3. a)..., b)...

----

The order in which the references are grouped together as a) b) and c) in each composite reference is simply determined by the order in which they were added to the text. EndNote allows the user to "enter" each reference (e.g. by clicking on "[1]") and re-order the references as to which is a) and which is b).

That's it.

You can imagine a further theoretical "complication". Imagine a fourth reference is written. This should be able to borrow the numbering from the previous references. E.g.

"Various calculations have been performed to investigate such reactions."[1b-c, 3b]

References do not have to be grouped by author, or be ordered within the group by year. They are ordered the way they are put in.

That's all there is to it. It is perhaps why the "rules" aren't documented. They are simple and generally quite intuitive, the ordering is arbitrary and up to the author to choose.

---

THE LIMITATION IS SURPRISING

I completely understand that making such a program highly functional is a great effort on the part of the participants and that resources are scant.

That notwithstanding, this way of citing is far from rare if you consider that chemists number in the hundreds of thousands. Moreover, chemistry is the central science, and therefore this citation style spills over to journals that are, effectively, physical or biological in the topics they cover. So Zotero has a massive blind spot right in the centre of science.

Finally, it seems to me that in undertaking the creation of something like Zotero, there should be some awareness of such citation styles long before any code is written. We are talking here about a style of output that orders text in a particular way, and which can be seen clearly in virtually every chemical, biochemical, materials, theoretical and physical chemistry journal (probably well over 150 journals). From the discussions above it seems to me that one objection is that the actual "engine room" of the software doesn't permit sufficient flexibility to make this happen. That seems a bizarre limitation in 2016, not least since other software (EndNote for example) has managed this a long time ago, as someone already noted.

I understand it takes lots of effort and I am a chemist, not a programmer. But at the same time, it is a question of ordering text for a program whose function is to order text. In 2016.

@lavoisier, if you're interested, you could certainly create some momentum by helping us determine how exactly these styles work, and what kind of variants exist. There are a bunch of additional questions that need answering beyond the information you gave above. I'd be happy to take care of the "engine room" end of things, so no programming experience required on your part.