Tests ----- http://www.gtalbot.org/BrowserBugsSection/CSS3WritingModes/draft-multi-column-vrl-002.xht http://www.gtalbot.org/BrowserBugsSection/CSS3WritingModes/draft-multi-column-vlr-003.xht http://www.gtalbot.org/BrowserBugsSection/CSS3WritingModes/draft-multi-column-vrl-006.xht http://www.gtalbot.org/BrowserBugsSection/CSS3WritingModes/draft-multi-column-vlr-007.xht Expected -------- http://www.gtalbot.org/BrowserBugsSection/CSS3WritingModes/multicol-count-002-ref.xht Explanations ------------ I am not entirely sure that this is a bug. That's why I do not mark this bug report as NEW. If 'height' (logical width) is set instead of 'width' (logical height), then Firefox passes those tests. Eg. In draft-multi-column-vrl-006 , if you replace width: 21em; with height: 8em; then Firefox will render the expected. Notes ----- - Chrome 50.0.2661.94 passes all 4 tests - I am using Firefox 49.0a1 buildID=20160430030223 - I use Linux 3.13.0-85-generic x86_64, Qt: 4.8.6, KDE 4.13.3; Kubuntu (trusty) 14.04.4 LTS - I've searched for duplicates and did not find any.

7.3.2. Auto-sizing Block Containers in Orthogonal Flows http://www.w3.org/TR/css-writing-modes-3/#auto-multicol states " If a box establishing an orthogonal flow is either a block container or a multi-column element, for the case where the box’s inline size is auto: If column-count is not auto and column-width is auto, calculate the used column-width with the shrink-to-fit formula min(max-content, max(min-content, (column-count − 1) × column-gap)), where: min-content the min-content inline-size of the box max-content the max-content inline-size of the box " min-content and max-content are 1em in tests 2 and 3; min-content and max-content are 4em in tests 6 and 7; 'width:19em' and 'width: 21em' (now commented in the tests) (logical height or block size) should not even be in those tests in order to be pure, better challenges. Marking as NEW David, thanks for the better summary.

> 'width:19em' and 'width: 21em' (now commented in the tests) I am putting those width declarations back, otherwise the summary will no longer make sense :) But I am convinced those tests would be better challenges without those width declarations. I will create some variations of those tests where those width declarations will be absent.

Additional tests where 'width' of the multi-column element has been deliberately set to 'auto': http://www.gtalbot.org/BrowserBugsSection/CSS3WritingModes/draft-multi-column-vrl-010.xht http://www.gtalbot.org/BrowserBugsSection/CSS3WritingModes/draft-multi-column-vlr-011.xht http://www.gtalbot.org/BrowserBugsSection/CSS3WritingModes/draft-multi-column-vrl-014.xht http://www.gtalbot.org/BrowserBugsSection/CSS3WritingModes/draft-multi-column-vlr-015.xht Chrome 50.0.2661.94 also passes those additional tests.