At Mathematical Reviews, we work hard to make sure that our bibliographic data are correct. We have more than twenty people in our Acquisitions and Cataloging Departments who verify, verify, verify. So that you don’t have to repeat our work, we try to make it easy for users of MathSciNet^{®} to obtain proper citations. The most popular method is to choose **BibTeX** as the “Alternative Format”.

### Saving a result as a citation

If you want to save the citation to just one item, here is an example that shows you how to do that. Let’s say that after a search, we come across the paper by Mumford and Shah where they define the Mumford-Shah functional, which is useful in computer vision, materials science, solid mechanics, and other areas where variational methods are important.

In the top left, we see that we can select an alternative format. Here are the choices:

Selecting **BibTeX** produces:

@article {MR997568, AUTHOR = {Mumford, David and Shah, Jayant}, TITLE = {Optimal approximations by piecewise smooth functions and associated variational problems}, JOURNAL = {Comm. Pure Appl. Math.}, FJOURNAL = {Communications on Pure and Applied Mathematics}, VOLUME = {42}, YEAR = {1989}, NUMBER = {5}, PAGES = {577--685}, ISSN = {0010-3640}, CODEN = {CPAMA}, MRCLASS = {49F20 (41A30 92A27)}, MRNUMBER = {997568 (90g:49033)}, MRREVIEWER = {Constantin Udri{\c{s}}te}, DOI = {10.1002/cpa.3160420503}, URL = {http://dx.doi.org/10.1002/cpa.3160420503}, }

You can now copy and paste this into your .bib file. Note that the BibTeX key is the MR number (MR997568). Most people switch that to something that can be remembered more easily.

### Saving multiple citations

Often, you may want to save a collection of citations for items on a particular theme. For instance, if you are writing a paper about the Mumford-Shah conjecture, you may want to have citations for lots of papers related to the original. In the top right-hand corner,

we can see that there are 343 references to the Mumford-Shah paper from reference lists and 23 references to the paper from a review (i.e., the reviewer of another paper explicitly mentions the Mumford-Shah paper). Assuming that the instances from the reviews are a good core sample, let’s use those. Clicking on “From Reviews” gives us the list. My preferences are set to display only 20 items per page, so we want to click

to have all 23 on screen. Now, at the top left we see

Clicking the box with **Reviews (HTML)** brings up the choices:

Let’s pick BibTeX again

Now you can check the boxes for your favorites among the collection to mark them, then retrieve just this subset of citations. Or you can pick “Retrieve First 50” and retrieve all 23 citations (since 23 ≤ 50), all formatted for us in BibTeX. Here are the first two:

@article {MR3105370, AUTHOR = {Zhang, Ruiliang and Bresson, Xavier and Chan, Tony F. and Tai, Xue-Cheng}, TITLE = {Four color theorem and convex relaxation for image segmentation with any number of regions}, JOURNAL = {Inverse Probl. Imaging}, FJOURNAL = {Inverse Problems and Imaging}, VOLUME = {7}, YEAR = {2013}, NUMBER = {3}, PAGES = {1099--1113}, ISSN = {1930-8337}, MRCLASS = {94A08 (52B55 65K10 68U10)}, MRNUMBER = {3105370}, MRREVIEWER = {Mikl{\'o}s Hoffmann}, DOI = {10.3934/ipi.2013.7.1099}, URL = {http://dx.doi.org/10.3934/ipi.2013.7.1099}, }

@incollection {MR3057536, AUTHOR = {Focardi, Matteo}, TITLE = {Regularity issues for local minimizers of the {M}umford \& {S}hah energy in 2{D}}, BOOKTITLE = {Bruno {P}ini {M}athematical {A}nalysis {S}eminar 2012}, SERIES = {Bruno Pini Math. Anal. Semin.}, VOLUME = {2012}, PAGES = {14--32}, PUBLISHER = {Univ. Bologna, Alma Mater Stud., Bologna}, YEAR = {2012}, MRCLASS = {49N60 (49K10)}, MRNUMBER = {3057536}, MRREVIEWER = {Michela Eleuteri}, }

For more information, see the MathSciNet help page on Retrieving Citations.

### References from the command line

Andrew Comech, from Texas A&M, wrote a command-line script for retrieving references in BibTeX format. It requires a subscription to MathSciNet, but it serves the needs of those who prefer not to point and click repeatedly. The script is bibget, and is available from Comech’s webpage: http://www.math.tamu.edu/~comech/tools/bibget/bibget. The script has an option that allows you to use an SSH tunnel, so that you can use the tool from off campus. This can fail (for the usual unknown reasons). As an alternative, using Comech’s example, you can try `ssh <your host name> bibget a=gilkey t=invariance book 1984 2>/dev/null`

Notes: This is a `bash`

script, so you will want to use it on a linux or unix machine (or from a terminal window on a Mac). Or you can load a port of `bash`

for Windows on your machine. It also uses `wget`

or `lynx`

to reach the network, so make sure you have one of those installed. For you Mac users, bibget supports `curl`

, which is built into most Macs. Comech tells me that the script is “poorly coded’, but it does get the job done. I have met some mathematicians who use bibget to build up a master BibTeX file, which they then use for all their papers.

### References without a subscription

There is a tool to verify references that works even if you do not have a subscription to MathSciNet handy. It is called **mref.** If you type (or copy and paste) the author+title of a paper (or author + title + some journal info, if there is no unique match), you get back our bibliographic listing of the item, including volume number, pages, year, and the MR number. The MR number comes with a live link to the item in MathSciNet. For instance, the search

Andrew Comech Optimal regularity of Fourier integral operators with one-sided folds. Comm. Partial Differential Equations

produces

Comech, Andrew. Optimal regularity of Fourier integral operators with one-sided folds. Comm. Partial Differential Equations 24 (1999), no. 7-8, 1263–1281. MR1697488 (2000m:35190).

The `mref`

search can be a little fussy. It is usually best to include information about the author, title, and the journal. We developed the tool originally for authors and publishers to have a way of verifying bibliographic information in reference lists for journal articles and books they were working on. Having the live link to the item in MathSciNet is a feature of reference lists in AMS online publications. ~~Note that ~~ `mref`

does not offer you the option of BibTeX format.**Correction**: Yes it does! There is a big “BibTeX” button underneath the search window.

A big thanks to Mathematical Reviews for producing clean and reliable bib entries. Oddly enough, the bib entries provided by publishers often require quite some post-processing.

The bibget makes my life easier. Thank you for telling us about it.

I’d also like to make something similar but I hesitate to do it because of the ‘Copyright Information and Terms of Use’, which says you may not conduct automated searching or downloading, by use of scripted searches, robots, spiders, crawlers, or otherwise. To what extent are we able to ‘conduct automated searching’?

Thank you for your message about MathSciNet and information about BibTeX and bibget.

We prohibit most scripted searches because of the need to abide by the terms of our agreements with publishers and to safeguard the database itself. We don’t object to researchers using MathSciNet to create the reference list / bibliography for a paper or for organizing their own CVs. Indeed, we have some tools for batch downloads that help in that regard. (See: http://www.ams.org/publications/math-reviews/tools/tools.) In that direction, we are working on creating a tool (on our site) similar to bibget that doesn’t conflict with our concerns about automated searches. We hope to have it by the end of the year.

> We hope to have it by the end of the year.

Any news on this tool? An AMS-blessed solution would be very welcome, and could be much more sensitive to what you regard as reasonable use than an outside tool.

There is no progress to report here.

On the other hand, I did discover that BibDesk, which comes with TeXShop for Macs is an effective tool for references.

I need to cite a review to a certain article (the reason is that the review contains an idea which is not in the article). I try to adapt incollection, but may be it is not the best way

If you want to add something to a reference list, the form would be:

Kenneth A. Ribet, review of “On the conjecture of Birch and Swinnerton-Dyer,” (in Invent. Math. 39 (1977), no. 3, 223–251, by J. Coates and A. Wiles), Mathematical Reviews/MathSciNet MR0463176 http://www.ams.org/mathscinet-getitem?mr=0463176

If you want to have something in text, hyperlink the MR number “MR0463176” using the URL http://www.ams.org/mathscinet-getitem?mr=0463176 . In the prose surrounding the reference, try to mention the reviewer’s name and “Mathematical Reviews” in full (avoid phrasing such as “Math Reviews” or “the Mathematical review”).

Example:

K. Ribet (see his review published in Mathematical Reviews [MR0463176])

states that the hypothesis that the class number is 1 is not essential in the result of Coates and Wiles, referring to work of N. Arthaud.