CSS Regions Module Level 1

W3C Working Draft, 9 October 2014

This version:
Latest version:
Editor's Draft:
Previous Versions:
www-style@w3.org with subject line “[css-regions] … message topic …” (archives)
Test Suite:
(Microsoft Corporation)
(Adobe Systems, Inc.)
Former Editors:
(Adobe Systems, Inc.)
Issues list:
In Bugzilla


The CSS Regions module allows content from one or more elements to flow through one or more boxes called CSS Regions, fragmented as defined in [CSS3-BREAK]. This module also defines CSSOM to expose both the inputs and outputs of this fragmentation.

CSS is a language for describing the rendering of structured documents (such as HTML and XML) on screen, on paper, in speech, etc.

Status of this document

This section describes the status of this document at the time of its publication. Other documents may supersede this document. A list of current W3C publications and the latest revision of this technical report can be found in the W3C technical reports index at http://www.w3.org/TR/.

Publication as a Working Draft does not imply endorsement by the W3C Membership. This is a draft document and may be updated, replaced or obsoleted by other documents at any time. It is inappropriate to cite this document as other than work in progress.

The (archived) public mailing list www-style@w3.org (see instructions) is preferred for discussion of this specification. When sending e-mail, please put the text “css-regions” in the subject, preferably like this: “[css-regions] …summary of comment…

This document was produced by the CSS Working Group (part of the Style Activity).

This document was produced by a group operating under the 5 February 2004 W3C Patent Policy. W3C maintains a public list of any patent disclosures made in connection with the deliverables of the group; that page also includes instructions for disclosing a patent. An individual who has actual knowledge of a patent which the individual believes contains Essential Claim(s) must disclose the information in accordance with section 6 of the W3C Patent Policy.

This document is governed by the 1 August 2014 W3C Process Document.

Table of Contents

1. Introduction

This section is non-normative.

The core concept behind CSS Regions is the ability to say, "Display this content (a named flow) over there (a region chain)." The simplest example is:

  #this {
    flow-into: my-flow;
  #there {
    flow-from: my-flow;

These two declarations will take the element that matches #this, put it into a flow named "my-flow", and display the contents of "my-flow" in the box from the element that matches #there. This example has a single content source for the named flow, and a single box for the region chain. Named flows can also have multiple sources and use multiple boxes for the region chain.

The named flow mechanism can be used in several different ways - some of which are custom overflow handling, aggregating content, linked display boxes, magazine-style layout, and flowing content through areas in a paginated view.

Linked display boxes can be created to display article content above and below other content on a page. Given markup like this:

  <article> ...some content... </article>
  <aside> ad or image content </aside>

The <aside> content will be displayed below all of the article content. On a large screen the page might display without scrolling, but on a small screen the <aside> content might not be visible until the view scrolls. If it’s important to show at least some of the <aside> content in the initial view, CSS Regions can fragment the article content across two boxes - one above the <aside> and one below.

In this example (for simplicity’s sake) we create the boxes with additional elements:

  <article> ...some content... </article>
  <div class="top region"></div>
  <aside> ad or image content </aside>
  <div class="region"></div>

  article {
    flow-into: article-flow;
  .region {
    flow-from: article-flow;
  .top {
    max-height: 80vh;

So the top box in the region chain is limited to 80% of the viewport height. If the article content doesn’t fit in that box, the article will continue in the second box after the <aside> content.

Article and aside rendering without CSS Regions
Rendering without CSS Regions
Article and aside rendering with CSS Regions
Rendering with CSS Regions

In the images above, the gray area represents the content below the screen edge in the initial view.

This example links just two boxes together, but more boxes could be added to the region chain to regularly interleave other content with the article.

Custom overflow handling can be accomplished by linking a separate overflow box. In this example, the overflow box is nestled inside a menu in the markup, and only displays if the menu is toggled.

  <nav> ...some links... </nav>
  <div class="menu">
    ...some more links...

If the links in the main nav element are placed in a named flow, that flow can be directed through both the main nav element and the overflow nav box in the menu:

  nav a {
    flow-into: nav-link-flow;
  nav {
    flow-from: nav-link-flow;

Then the main nav element and the menu can be arranged with constraints such that when the screen is too narrow for the main nav element to display all of the navigation links, the overflow moves to the menu.

Wide nav bar showing all of the links
Wide rendering with menu shown
Narrow nav bar with some of the links in the menu
Narrow rendering with menu shown

Since content is assigned to a named flow using a CSS selector, the content can come from multiple sources. The resulting aggregation can be displayed in a single box or flowed through multiple boxes as above.

So given this markup:

  <div class="breaking-news"></div>
  <article>News story</article>
  <article class="breaking">Sports story</article>
  <article>News story</article>
  <article class="breaking">Entertainment story</article>
  <article>Sports story</article>

You can take the "breaking" stories and display them above all the others using two lines of CSS:

  .breaking {
    flow-into: breaking-news;
  .breaking-news {
    flow-from: breaking-news;

Given more data accessible to CSS selectors, you could rearrange the articles in other ways (sports on top, etc.) depending on the user’s preferences.


CSS Regions are independent from layout

Any of the CSS layout facilities can be used to create, position and size boxes that can become CSS Regions.

The CSS Regions module does not define a layout mechanism and is meant to integrate with existing and future CSS layout facilities.


CSS Regions do not have to be elements

The CSS Regions module is independent of the layout of boxes and the mechanisms used to create them.

For simplicity, our examples tend to use elements to define the boxes. Any other mechanism available in markup or style to create stylable boxes could be used instead, such as pseudo-elements or the @slot rule proposed by the CSS Page Template Module [CSS3-PAGE-TEMPLATE].

The only requirement for box to become a CSS Region is that the flow-from property applies to the box.

2. CSS Regions concepts

2.1. Regions

A CSS Region is a block container that has an associated named flow (see the flow-from property).

2.2. Region chain

A region chain is the sequence of regions that are associated with a named flow. CSS Regions in a region chain receive content from the named flow following their order in the chain. CSS Regions are organized into a region chain according to the document order.

2.3. Named flows

A named flow is the ordered sequence of content associated with a flow with a given identifier. Contents in a named flow are ordered according to the document order.

Content is placed into a named flow with the flow-into property. The content in a named flow is laid out in the region chain that is associated with this named flow using the flow-from property.

Content from a named flow is broken up between regions according to the regions flow breaking rules.

A named flow is created when some content is moved into the flow with the given identifier or when at least one CSS Region requests content from that flow.

2.4. Regions flow breaking rules

Breaking a named flow across a region chain is similar to breaking a document’s content across pages (see [CSS3PAGE]) or a multi-column element’s content across column boxes (see [CSS3COL]). One difference is that page boxes are generated based on the available content whereas a region chain is a set of recipient boxes for the named flow content whose dynamic generation is not in the scope of this specification.

Each CSS Region in turn consumes content from its associated named flow. The named flow content is positioned in the current region until a natural or forced region break occurs, at which point the next region in the region chain becomes the current region. If there are no more CSS Regions in the region chain and there is still content in the flow, the positioning of the remaining content is controlled by the region-fragment property on the last CSS Region in the chain.

The CSS regions module follows the fragmentation rules defined in the CSS Fragmentation Module Level 3 (see [CSS3-BREAK]).

3. Properties

3.1. The flow-into property


[Shadow]: getFlowByName and shadow DOM

The ‘flow-into’ property can place an element or its contents into a named flow. Content that belongs to the same flow is laid out in the region chain associated with that flow.

Value:none | <ident> [element|content]?
Applies to:All elements, but not pseudo-elements such as ::first-line, ::first-letter, ::before or ::after.
Computed value:as specified
The element is not moved to a named flow and normal CSS processing takes place.
If the keyword element is present or neither keyword is present, then the element is taken out of its parent’s flow and placed into the flow with the name '<ident>'. If the keyword content is present, then only the element’s contents are placed into the named flow. The element or content is said to have a specified flow. The values none, inherit, default, auto and initial are invalid flow names.

The flow-into property affects the visual formatting of elements or contents placed into a named flow and of the region chain laying out content from a named flow. The flow-into property does not affect the CSS cascade and inheritance for the elements on which it is specified. The flow-into property does not affect the DOM [DOM] position of an element or its contents. The flow-into property does not affect ordering in non-visual media (such as speech). Likewise, flow-into does not affect the default traversal order of sequential navigation modes (such as cycling through links, see e.g. nav-index [CSS3UI] or tabindex [HTML40]).

A named flow needs to be associated with a region chain (one or more CSS Regions) for its content to be visually formatted. If no region chain is associated with a given named flow, the content in the named flow is not rendered: it does not generate boxes and is not displayed.

The children of an element or content with a specified flow may themselves have a specified flow, in which case they become the next sibling of the latest element or content collected in that flow. In some cases, the child can become the next sibling for one of its ancestors in the same flow.

Content in a named flow is sequenced in document order. The visual formatting model uses the relationships between content in the named flow as input, rather than the contents’ position in the DOM.

Each CSS region in a region chain establishes a containing block for absolutely positioned elements in the named flow (see [CSS21]). That first CSS region in a region chain establishes the initial containing block for such absolutely positioned elements. Regions don’t establish a containing block for fixed positioned elements in the named flow. Such fixed positioned elements are still positioned relative to the viewport or the page area even if they have been redirected into a named flow

The first region defines the principal writing mode for the entire flow. The writing mode on subsequent regions is ignored.


The flow-into property moves an element into the flow and the interplay with selectors should be considered carefully.

For example,

table {flow-into: table-content}

will move all tables in the "table-content" named flow. However, the

table > * {flow-into: table-content} ...

selector will move all immediate children of all table elements into the "table-content" named flow (which may be useful as it will usually result, if the immediate children are rows, in merging rows of multiple tables), but the

table * {flow-into: table-content}

selector will move all descendants of table elements into the "table-content" named flow, transforming the element tree into a flat list in order of opening tags (which is probably not intentional). This will make all the descendants of table elements siblings in the named flow. Having the descendants become siblings in the named flow results in a different processing (see CSS 2.1’s anonymous table objects). This note illustrates how authors must exercise caution when choosing a particular selector for setting the flow-into property to avoid unintended results.


Another consequence of moving elements into named flows is that surrounding whitespace is not moved into the named flow. If you have code like this:

span {flow-into: span-content}

Then the "span-content" named flow contents will contain this:


Which will change the display from "one two" to "onetwo". If whitespace is significant, then moving the parent that contains the whitespace to the named flow is required.

3.2. The flow-from property

The flow-from property makes a block container a region and associates it with a named flow.

Value:<ident> | none | inherit
Applies to:Non-replaced block containers.
This might be expanded in future versions of the specification to allow other types of containers to receive flow content.
Computed value:as specified
The block container is not a CSS Region.
The block container becomes a CSS Region (except as detailed in the text below), and is ordered in a region chain according to its document order. The content from the flow with the <ident> name will be broken into fragments and visually formatted in the principal boxes of the regions in the region chain.
If there is no flow with name <ident>, then the block container does not format any content visually.

If the content property computes to something else than normal (or none for a pseudo-element), the block container does not become a CSS Region. If the display property of the block container or one of its ancestors computes to none, the block container does not become a CSS Region.

A CSS Region’s document children are not visually formatted unless they are directed to a named flow with an associated region chain.

Block container pseudo-elements where the value of flow-from computes to an <ident> and the value of content computes to none are generated as CSS Regions, which is an update to the behavior described in [CSS21].


A block container becomes a CSS Region when its flow-from property is set to a valid <ident> value, even if there is no content contributing to the referenced flow. For example:

    flow-into: thread;
    flow-from: thread;
  <div class=region>div content</div>

There is no element matching the .article selector and therefore no content in the thread flow. However, the block container matching the .region selector is still associated with that empty named flow and, consequently, its children are not formatted visually.


At the time of this note-writing, the display values that always result in a non-replaced block container include block, inline-block, table-cell, table-caption, and list-item. All of these display values work as regions with non-replaced elements.

The flex and grid display values do not result in block containers (they are defined as flex containers and grid elements, respectively). So ‘flow-from’ combined with those display values does not result in a CSS Region.

CSS Regions create a new stacking context. CSS Regions establish a new block formatting Context. Exclusions (see [CSS3-EXCLUSIONS]) potentially impact the content laid out in region chains, just as for non-regions.


With region chains, an element may be split across multiple boxes and these boxes may overlap (for example if they are absolutely positioned). So fragments of the same element can overlap each other. Since each element has a single z-index, it would be required to find another mechanism to decide in which order the fragments are rendered. Since each CSS Region creates a new stacking context, it is clear that each fragment is rendered separately and their rendering order follows the regular CSS rendering model.

Fragments rendering separately is also relevant to elements that might normally be rendered as a unit (for example, an element with its own stacking context, or with transparency). Each fragment of these elements is separately contained in the stacking context created by the CSS Region, so each fragment of these elements is rendered separately.

See the regions visual formatting details section for a description of how width and height values are resolved for CSS Region boxes.

3.2.1. Cycle Detection

named flows containing elements where the value of flow-from computes to an <ident> can produce nonsensical circular relationships, such as a named flow containing CSS Regions in its own region chain. These relationships can be easily and reliably detected and resolved, however, by keeping track of a dependency graph and using common cycle-detection algorithms.

The dependency graph consists of edges such that:

If the graph contains a cycle, any elements where the value of flow-from computes to an <ident> participating in the cycle do not become CSS Regions.


For example, styling like this:

#id {
  flow-into: foolish;
  flow-from: foolish;

would move the #id element to a "foolish" named flow, and try to make the #id element a CSS Region for the "foolish" named flow. The "foolish" named flow would then contain its own region, creating a cycle. So the #id element does not become a CSS Region.

3.2.2. Nested fragmentation contexts

A CSS Region that contains a fragment of a named flow can itself be fragmented if it is nested within a fragmentation context [CSS3-BREAK], such as when a layout using a region chain is printed. In these cases break opportunities in the named flow fragment contained by the CSS Region are determined using the standard fragmentation rules. In other words, each region box and its associated fragment should break as if it were a simple div containing the fragment contents. This can be controlled by using an avoid break value on the CSS Region, if that is desired.

A CSS Region can be part of the contents of a separate named flow, as long as there are no cycles broken by the Cycle Detection described above. This case is a nested region context, which has an effect on the Visual Formatting Steps described below.

3.3. Region flow break properties: break-before, break-after, break-inside


This section is also defined in [CSS3-BREAK]. If that specification moves to last call with the region values, the section here can be replaced by a reference.

User agents laying out content in multiple regions must determine where content breaks occur. The problem of breaking content into fragments fitting in regions is similar to breaking content into pages or columns.

Each break ends layout in the current region and causes remaining pieces of content from the named flow to be visually formatted in the following region in the region chain, if there is one.

The following extends the break-before, break-after and break-inside properties from the [CSS3COL] specification to account for regions. The additional values are described below.

Value:auto | always | avoid | left | right | page | column | region | avoid-page | avoid-column | avoid-region
Applies to:block-level elements
Computed value:as specified
Value:auto | always | avoid | left | right | page | column | region | avoid-page | avoid-column | avoid-region
Applies to:block-level elements
Computed value:as specified
Value:auto | avoid | avoid-page | avoid-column | avoid-region
Applies to:block-level elements
Computed value:as specified

These properties describe page, column and region break behavior before/after/inside the generated box. These values are normatively defined in [CSS3COL]:

This specification adds the following new values:

Always force a region break before (after) the generated box.
Avoid a region break before (after, inside) the generated box.

The behavior of region breaks with regards to regions is identical to the behavior of page breaks with regards to pages, as defined in the [CSS21].

3.4. The region-fragment property

Value:auto | break
Applies to:CSS Regions
Computed value:as specified

The region-fragment property controls the behavior of the last region associated with a named flow.

Content flows as it would in a regular content box. If the content exceeds the container box, it is subject to the overflow property’s computed value on the CSS Region. Region breaks must be ignored on the last region.

If the content fits within the CSS Region, then this property has no effect. If the content does not fit within the CSS Region, the content breaks as if flow was going to continue in a subsequent region. See the breaking rules section. A forced region break takes precedence over a natural break point.

Flow content that follows the last break in the last region is not rendered.

The region-fragment property does not influence the size of the region it applies to.

The following code sample illustrates the usage of the region-fragment property.

  article {
    flow-into: article-flow;
  #region-1, #region-2 {
    flow-from: article-flow;
    region-fragment: break; /* or auto */
    overflow: visible; /* or hidden */
  <div id="region-1"></div>
  <div id="region-2"></div>
article with two
overflowing lines
region-fragment: break
overflow: visible
region-fragment: auto
overflow: visible
flow content rendering rendering with region-fragment:break and overflow:visible rendering with region-fragment:auto and overflow:visible
region-fragment: break
overflow: hidden
region-fragment: auto
overflow: hidden
rendering with region-fragment:break and overflow:hidden rendering with region-fragment:auto and overflow:hidden
Combinations of region-fragment and overflow.

The overflow property is honored on a region: if region content overflows, such as the borders of elements on the last line, the overflow property controls the visibility of the overflowing content. See the overflow property definition ([CSS21]).


Since content may flow into multiple regions, authors need a way to determine if there are enough regions to flow all the content from a named flow. This is especially important considering that the size of regions or the size of the content may change depending on the display context. For example, flowing the same content on a mobile phone with a small screen may require more regions than on a large desktop display. Another example is the user changing the font size of text flowing through regions. Depending on the change, new regions may be needed to accommodate for the additional space required to fit the larger text or some regions may need to be removed for smaller text.

4.1. The NamedFlow interface

The following APIs allow scripts to reference a NamedFlow object representation of a named flow.

An additional attribute on the Document interface provide access to named flows.

partial interface Document {
  readonly attribute NamedFlowMap namedFlows;

The namedFlows attribute on the Document interface returns a static snapshot of all the current named flows in the document.

The namedFlows map must include all named flows that are currently in the CREATED state. The list must not include named flows that are in the NULL state.

The NamedFlowMap interface provides a map of current NamedFlow instances in the document. The NamedFlowMap object is a snapshot of the data, and is read-only.

[MapClass(DOMString, NamedFlow)] interface NamedFlowMap {
  NamedFlow? get(DOMString flowName);
  boolean has(DOMString flowName);
  NamedFlowMap set(DOMString flowName, NamedFlow flowValue);
  boolean delete(DOMString flowName);

The map entries in a NamedFlowMap object are the named flow idents paired with their NamedFlow objects. The get() and has() methods return null and false respectively if there is no NamedFlow with the given ident. The set() and delete() methods always throw an InvalidAccessError exception, as this map is read-only. The NamedFlowMap interface uses the rest of the default map class methods.

The NamedFlow interface offers a representation of a named flow instance.

The NamedFlow interface can be used for different purposes. For example, the getRegionsByContent() method can help navigate by bookmark: a script can find the CSS Regions that display a particular anchor and bring them to view.

Likewise, the interface allows authors to check if all content has been fitted into existing regions. If it has, the overset attribute would be false.

interface NamedFlow : EventTarget {
  readonly attribute DOMString name;
  readonly attribute boolean overset;
  sequence<Region> getRegions();
  readonly attribute short firstEmptyRegionIndex;
  sequence<Node> getContent();
  sequence<Region> getRegionsByContent(Node node);

The name attribute returns the name of the NamedFlow instance.

The overset attribute returns true if there are named flow fragments that do not fit in the associated region chain (including the case where there are named flow fragments but no regions in the region chain). Otherwise, it returns false.

The getRegions() method returns the sequence of regions in the region chain associated with the named flow. Note that the returned values is a static sequence in document order.

The firstEmptyRegionIndex is the index of the first region in the region chain with the regionOverset attribute set to empty. If all regions have the regionOverset attribute set to fit or overset, the value for firstEmptyRegionIndex is -1. If there are no regions in the region chain, the value is -1 as well.

The getContent() method returns an ordered collection of nodes that constitute the named flow. The returned list is a static snapshot of the named flow content at the time the method is invoked. This list contains the contents that were moved to the named flow by the flow-into property but not any descendants (unless the descendants are themselves moved to the named flow).

The getRegionsByContent() method returns the sequence of regions that contain at least part of the target content node if it belongs to the named flow directly or one of its ancestors belongs to the named flow. Otherwise, the method returns an empty sequence. The returned value is a static sequence in document order.

The named flow states are :

A NamedFlow object is live: it always represents the named flow with the corresponding name even if that named flow has transitioned to the NULL state.

4.2. The Region interface

The Region interface is a supplemental interface which must be implemented by all objects ( Elements, pseudo-elements or other CSS constructs such as slots) in an implementation which can be CSS Regions.

interface Region {
  readonly attribute DOMString regionOverset;
  sequence<Range>? getRegionFlowRanges();
Element implements Region;

The regionOverset attribute returns one of the following values:

The region is the last one in the region chain and not able to fit the remaining content from the named flow. Note that the region’s overflow property value can be used to control the visibility of the overflowing content and the region-fragment property controls whether or not fragmentation happens on the content that overflows the last region.
The region’s flow fragment content fits into the region’s content box. If the region is the last one in the region chain, it means that the content fits without overflowing. If the region is not the last one in the region chain, that means the named flow content may be further fitted in subsequent regions. In this last case, note that the named flow fragment may be empty (for example if the region is too small to accommodate any content). This value is returned if the Region object is not (or no longer) a region.
All content from the named flow was fitted in prior regions.

If there is no content in the named flow, all regions associated with that named flow should have their regionOverset attribute return empty. If there is content in the flow but that content does not generate any box for visual formatting, the overset attribute on the first region in the region chain associated with the flow will return fit.

The getRegionFlowRanges() method returns an array of Range instances corresponding to fragment from the named flow that is laid out in the region. If the region has not received a fragment because it is too small to accommodate any, the method returns a single Range where the startContainer and startOffset have the same values as the endContainer and endOffset and therefore the collapsed attribute on the Range is true. In that situation, if the region is the first in the region chain, the startContainer is the first Node in the named flow and the startOffset is zero. If the region is the last region in the region chain (but not the first and only one), the startContainer and startOffset are the same values as the endContainer and endOffset on the previous region in the region chain. The method returns null if the region object is not (or no longer) a region.

A Region instance may represent an object that is no longer a region. This may happen for example if the flow-from property on the corresponding pseudo-element, element or other construct becomes none but a script is still holding a reference to the Region object.

4.3. Named flow events

NamedFlow objects are EventTargets which dispatch the following events for their respective triggers. These events are asynchronous, and fire at the end of the regions visual formatting steps.

The regionfragmentchange event is dispatched on any change to a named flow’s fragmentation through its region chain, including changes to any overset fragment.

Type regionfragmentchange
Interface UIEvent (see [DOM-LEVEL-3-EVENTS])
Sync / Async Async
Bubbles No
Target NamedFlow
Cancelable Yes
Default action none
Context info
  • Event.target: NamedFlow whose fragmentation has changed.

The regionoversetchange event is dispatched if any of the regionOverset values change in a named flow’s region chain, including when regions are added or removed from the chain.

Type regionoversetchange
Interface UIEvent (see [DOM-LEVEL-3-EVENTS])
Sync / Async Async
Bubbles No
Target NamedFlow
Cancelable Yes
Default action none
Context info
  • Event.target: NamedFlow whose region chain has regionOverset values that have changed.

4.4. Clarifications on pre-existing APIs

4.4.1. getClientRects() and getBoundingClientRect()

The CSSOM View Module defines how user agents compute the bounding client rectangle for an element (getBoundingClientRect()) and its generated boxes (getClientRects()).

This definition applies to the (possibly) multiple boxes generated for an element in a named flow flowing through a region chain. The getClientRects() method returns the list of boxes generated for each of the element fragments laid out in different regions. The getBoundingClientRect() method operates as specified in the CSSOM View Module as well and is computed from the set of rectangles returned by getClientRects().

4.4.2. offsetTop, offsetLeft, offsetWidth, offsetHeight and offsetParent

The computation of the offset attributes for elements laid out in a named flow follow the specification [CSSOM]. For the purpose of these algorithms, the first CSS layout box associated with an element laid out in a named flow is the first box generated for the first region the element is laid out into. In the offsetParent algorithm, the nearest ancestor search skips from the topmost named flow elements directly to the body element.

5. Multi-column regions

A multi-column [CSS3COL] element can be assigned to a region chain. The element becomes part of the region chain for the associated named flow, and flows its content fragments through columns according to the multi-column specification [CSS3COL]. In particular, when computing the flow fragment height of a multi-column element that is associated with a named flow, the column-fill [CSS3COL] property is honored to balance the fragments of content that would flow through its columns.

Overflow of multicol regions is mostly handled according to the same rules as other CSS Regions. If the remainder of the named flow does not fit in the multicol region, then the rest of the content flows into the remaining region chain. However, if a multicol region is the last region in a region chain, then the multicol region must follow the overflow column rules [CSS3COL].

The following example:

  article {
      flow-into: article-flow;
  #multi-col {
      column-count: 2;
      flow-from: article;
      height: 6em;
      column-gap: 1em; 
  #remainder {
      flow-from: article;
      height: auto;
  <div id="multicol"></div>
  <div id="remainder"></div>

is equivalent in rendering to, for example:

  article {
      flow-into: article-flow;
  #flex {
      display: flex;
      flex-pack: justify;
      height: 6em;
  #flex > div {
      flow-from: article;
      width: calc(50% - 0.5em);
  #remainder {
      flow-from: article;
      height: auto;
  <div id="flex">
    <div />
    <div />
  <div id="remainder"></div>

6. Pseudo-elements

It can be useful to visually mark the content to highlight that a content thread is flowing through the region chain. For example, a marker such as 'continued below' clearly indicates, at the end of a CSS Region, that there is more content in the flow which can be found by scrolling past whatever content interrupts the region chain.

The '::before' and '::after' pseudo-elements (see [SELECT]) let content authors mark the beginning and end of a region with such markers.

7. Regions visual formatting details

Regions are laid out by CSS and take part in the normal box model and other layout models offered by CSS modules such as flexible boxes ([CSS3-FLEXBOX]). However, regions lay out a fragment of their named flow instead of laying out descendant content as happens with other boxes.

This section describes the model for laying out regions and for laying out named flow content into regions. The descriptions in this section are biased towards a horizontal writing mode, using width for logical width (or measure) and height for logical height (or extent) as defined in the CSS Writing Modes Module [CSS3-WRITING-MODES]). To use this model in a vertical writing mode apply the principles described in that specification.

7.1. Processing model

The '::before' content is laid out in the region prior to any other content coming from the flow.

The '::after' content is laid out in the region after laying out the flow fragment content into the RFCB. Then, flow content is removed from the fragment to accommodate the '::after' content. Accommodating means that the '::after' content is laid out without overflowing the region.

If there is not enough room to accommodate the ::before content, the '::after' content after removing all flow fragment content, or a combination of the two, then the ::before and/or ::after content overflows that region.

7.2. The Region Flow Content Box (RFCB)

A region box lays out the following boxes:

The ::before, RFCB and ::after boxes contained in the Region Box
The Region Flow Content Box (RFCB)

Laying out a region box follows the same processing rules as for any other block container box.

The RFCB is a block container box with a computed width of auto and whose used height is resolved as detailed below.

Since the RFCB is a block container box, the ::before box and ::after box will also be block containers, though the contents of ::before and ::after may be inline within those boxes.

7.2.1. RFCB width resolution

At various points in the visual formatting of documents containing regions, the used width of RFCBs and regions need to be resolved. In all cases, the resolution is done following the rules for calculating widths and margins (see [CSS21]). Sometimes, resolving the used width value requires measuring the content’s min-content and max-content values (as defined in the CSS Writing Modes Module [CSS3-WRITING-MODES]). or an RFCB, these measures are made on the entire associated named flow content.

As a consequence, all RFCBs of regions associated with a given named flow share the same min-content and max-content measures.

This approach is consistent with the box model for breaking ([CSS3-BREAK]).

7.3. Regions visual formatting steps

Formatting documents that contain named flows laid out in regions is a three-step process:

visual representation of the three-step process
Regions visual formatting steps

7.3.1. Step 1: RFCB flow fragment height resolution

Conceptually, resolving the flow fragment height is a two phase process. RFCB flow fragment height resolution, Phase 1

The document is laid out with a used height of zero for all RFCBs. In this phase, the content of named flows is not laid out in regions. This phase yields resolved position and sizes for all regions and their RFCBs in the document. RFCB flow fragment height resolution, Phase 2

named flows are laid out in regions. The user agent resolves the flow fragment height for the RFCBs using the remainder of the flow and accounting for fragmentation rules. This process accounts for constraints such as the height or max-height values, as described in the CSS 2.1 section on calculating heights and margins (see the Block-level non-replaced elements in normal flow... section and the complicated cases section). During this phase, generated content is laid out according to the rules described earlier in this document.

In a nested region context, this phase will trigger the beginning of Step 1 for any inner named flows whose regions are contained in the outer named flow. All of Step 1 for inner flows must recursively complete before Step 1 for an outer flow completes.

7.3.2. Step 2: region boxes layout

In this step, the document is laid out according to the normal CSS layout rules.

If a measure of the content is required to resolve the used width of the region box, the value is resolved as described in the RFCB width resolution section.

If a measure of the content is required to resolve the used height of the RFCB (for example if the region box is absolutely positioned), the flow fragment height resolved in Step 1 is used for the vertical content measure of the RFCB.

At the end of this step, regions are laid out and ready to receive content from their associated named flows.

7.3.3. Step 3: named flows layout

In this final step, the content of named flows is laid out in the regions' RFCBs along with the generated content boxes.

The used width for RFCBs is resolved as described before.

The used height of RFCBs is resolved such that none of the boxes in the region box’s normal flow overflow the region’s box. In other words, the RFCB boxes are stretched vertically to accommodate as much of the flow as possible without overflowing the region box and accounting for fragmentation rules and generated content boxes.

During this phase, generated content is laid out according to the rules described earlier in this document.

In a nested region context, this step will trigger Step 2 for inner named flows whose regions are contained in the outer named flow. Fragmentation of the inner regions may result as they are laid out in the outer region chain. Once Step 3 for an outer named flow is complete, Step 3 for the inner named flows recursively begins.

Once Step 3 for a named flow is complete, The conditions for the named flow events are checked, and if the triggers are met the events dispatch at this point.


The model for resolving auto sized regions will cause, under certain circumstances, the flow content to be overset or underset. In other words, it will not fit tightly. The model prevents having circular dependencies in the layout model. Implementations may decide to apply further layout steps to ensure that the whole flow content is displayed to the user, even in edge cases.

7.4. Regions visual formatting: implementation note

The process for resolving an RFCB’s height and the three-step process used to lay out documents containing regions and named flows are conceptual descriptions of what the layout should yield and implementations should optimize to reduce the number of steps and phases necessary wherever possible.

7.5. Regions visual formatting example

This section is non-normative.

This example considers a document where content flows between three regions, and region boxes are intermixed with the normal document content.

  article {
      flow-into: article;
  #rA, #rB, #rC {
      flow-from: article;
      height: auto;
      margin: 1em 1em 0em 1em;
      padding: 1em;
      border: 3px solid #46A4E9;
  #rA {
      width: auto;
  #rB {
      float: left;
      width: 15em;
      max-height: 150px;
  #rC {
      float: right;
      width: 12em;
  #main-flow {
      padding: 0em 1em 0em 1em;
        <p style="break-after:region;">I am not a ... </p>
    <div id="rA"></div>
    <div id="rB"></div>
    <div id="rC"></div>
    <div id="main-flow">
        <p>Lorem ipsum dolor ...</p>

The following sections and figures illustrate the intermediate results for the visual formatting process. In the following, we call RFCB-A, RFCB-B and RFCB-C the RFCBs for regions rA, rB and rC respectively.

7.5.1. Step 1 - Phase 1: Laying out RFCBs with used height of zero

Applying the rules for Step 1, Phase 1, the computed auto width values for the RFCBs are resolved to used values according to the normal CSS layout rules meaning they stretch to the width of their containing block’s content box.

  1. RFCB-A: stretches to fit the rA content box.

    Since rA also has an auto width, its own used width is stretched to fit the <body> content box.

  2. RFCB-B: stretches to fit the rB content box.
  3. RFCB-C: stretches to fit the rC content box.

Also applying the rules for Step 1, Phase 1, the used values for the RFCBs height properties are all zero.

Conceptually, this produces the layout illustrated below.

Step 1 - Phase 1: Layout RFCBs with used heights of 0
Step 1 - Phase 1: Layout RFCBs with used heights of 0

7.5.2. Step 1 - Phase 2: Layout flow to compute the RFCBs' flow fragments heights

In this second phase of Step 1, the named flow is laid out in regions and the height of each fragment falling in each RFCB is computed.

The user agent lays out as much of the flow into an area with RFCB-A’s used width. rA’s height computes to auto and there is no vertical maximum height for RFCA’s height. However, because there is a break after the first paragraph in the "article" named flow, only this first paragraph is laid out in RFCB-A and FH-A (the flow fragment height for RFCB-A) is resolved by laying out this first paragraph in the used width.

At this point, the user agent lays out as much of the remaining flow as possible in RFCB-B. Because rB’s max-height computed value is "150px", the user agent only lays out the "article" named flow using RFCB-B’s used width until the point where laying out additional content would cause RFCB-B to overflow rB’s box. The fragment height for RFCB-B is resolved: FH-B (150px).

Finally, the user agent lays out the remainder of the flow in RFCB-C. Because rC has no other constraints and no region breaks, the remaining content is laid out in RFCB-C’s used width. This results in a resolved flow fragment height: FH-C.

Step 1 - Phase 2: Measure flow fragments heights
Step 1 - Phase 2: Measure flow fragments heights

7.5.3. Step 2: Layout document and regions without named flows

The used width of RFCB-A, RFCB-B and RFCB-C are resolved as in the previous step. However, the height is resolved differently.

Resolving the height of rA requires a content measure which is FH-A (the flow fragment height for RFCB-A).

The height of rB results from first computing its content measure and then applying the rules for max-height. Here, the vertical content measure evaluates to FH-B. After applying the rules for max-height and accounting for margins, padding and borders, the used height of rB is resolved to LH-B (150px).

The height of rC’s box results from calculating its content measure: FH-C becomes rC’s used height.

Step 2: Layout document and regions without named flows
Step 2: Layout document and regions without named flows

7.5.4. Step 3: named flows layout

In this final step, the article named flow is laid out in its region chain. The used width for each of the RFCB is resolved as in step 1 above.

The used height for the RFCB is a result of laying out the as much of the content in the region without overflowing its content box and following the fragmentation rules.

Because the computed width of the RFCB has not changed and the fragmentation rules applied are the same as in Phase 1, Step 2, the used height for RFCB-A, RFCB-B and RFCB-C are LH-A, LH-B and LH-C, respectively.

There may be situations where the used height of RFCBs resolved in Step 3 are different from the flow fragment height computed in Step 1 Phase 2.

Step 3: Final result after laying out named flows in regions
Step 3: Final result after laying out named flows in regions

8. Relation to document events

The CSS Regions module does not alter the normal processing of events in the document tree. In particular, if an event occurs on an element that is part of a named flow, the event’s bubble and capture phases happen following the document tree order.


This means that in most cases CSS Regions will not receive user events that trigger on their named flow content. Event handlers for named flow content can check getElementsFromPoint() [CSSOM-VIEW] to find the CSS Region where the user event occurred. Future versions of CSS-UI may provide a more general solution for user event bubbling where the stack of elements at the event coordinates

9. Relation to other specifications

This specification is related to other specifications as described in the references section. In addition, it is related to the following specifications:

  1. CSS Fragmentation Module Level 3 [CSS3-BREAK]. This module defines the rules for fragmenting content over multiple containers and applies to CSS Regions in addition to applying to multi-column and paged media.
  2. CSS Pagination Templates Module Level 3 [CSS3-PAGE-TEMPLATE]. This module defines a syntax to define layout templates which can be used when paginating content. The page templates use regions.
  3. CSS Exclusions Module [CSS3-EXCLUSIONS]. This module defines a generic way to define exclusions around which content can flow. This can be seen as an extension to CSS floats. In advanced layout designs, it is expected that the CSS Exclusions module will be commonly combined with the CSS Regions module.
  4. CSS Line Grid Module [CSS3-LINE-GRID]. This module defines a concept of line grid to align the position of lines in different elements. The line grid functionality is related and needed for aligning content flowing in separate regions.

10. Use Cases

Use cases are described on these pages.

11. Changes

11.1. Changes from February 18th 2014 version

11.2. Changes from May 28th 2013 version

11.3. Changes from August 28th 2012 version

11.4. Changes from May 3rd 2012 version

11.5. Older Changes

Older changelogs are archived on the CSSWG wiki.


The editors are grateful to the CSS working group for their feedback and help with the genesis of this specification.

In addition, the editors would like to thank the following individuals for their contributions, either during the conception of CSS Regions or during its development and specification review process:

Erik Arvidsson, Tab Atkins, Catalin Badea, Mihai Balan, Andrei Bucur, Razvan Caliman, Alexandru Chiculita, Phil Cupp, Arron Eicholz, John Jansen, CJ Gammon, Dimitri Glazkov, Daniel Glazman, Arno Gourdol, Catalin Grigoroscuta, David Hyatt, Brian Heuston, Ian Hickson, Jonathan Hoersch, Michael Jolson, Brad Kemper, Håkon Wium Lie, Kang-Hao (Kenny) Lu, Mihai Maerean, Markus Mielke, Robert O’Callahan, Edward O’Connor, Mihnea Ovidenie, Virgil Palanciuc, Olga Popiv, Christoph Päper, Anton Prowse, Peter Sorotokin, Elliott Sprehn, Radu Stavila, Christian Stockwell, Eugene Veselov, Boris Zbarsky, Stephen Zilles and the CSS Working Group members.


Document conventions

Conformance requirements are expressed with a combination of descriptive assertions and RFC 2119 terminology. The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in the normative parts of this document are to be interpreted as described in RFC 2119. However, for readability, these words do not appear in all uppercase letters in this specification.

All of the text of this specification is normative except sections explicitly marked as non-normative, examples, and notes. [RFC2119]

Examples in this specification are introduced with the words "for example" or are set apart from the normative text with class="example", like this:

This is an example of an informative example.

Informative notes begin with the word "Note" and are set apart from the normative text with class="note", like this:

Note, this is an informative note.

Advisements are normative sections styled to evoke special attention and are set apart from other normative text with <strong class="advisement">, like this: UAs MUST provide an accessible alternative.

Conformance classes

Conformance to this specification is defined for three conformance classes:

style sheet
A CSS style sheet.
A UA that interprets the semantics of a style sheet and renders documents that use them.
authoring tool
A UA that writes a style sheet.

A style sheet is conformant to this specification if all of its statements that use syntax defined in this module are valid according to the generic CSS grammar and the individual grammars of each feature defined in this module.

A renderer is conformant to this specification if, in addition to interpreting the style sheet as defined by the appropriate specifications, it supports all the features defined by this specification by parsing them correctly and rendering the document accordingly. However, the inability of a UA to correctly render a document due to limitations of the device does not make the UA non-conformant. (For example, a UA is not required to render color on a monochrome monitor.)

An authoring tool is conformant to this specification if it writes style sheets that are syntactically correct according to the generic CSS grammar and the individual grammars of each feature in this module, and meet all other conformance requirements of style sheets as described in this module.

Partial implementations

So that authors can exploit the forward-compatible parsing rules to assign fallback values, CSS renderers must treat as invalid (and ignore as appropriate) any at-rules, properties, property values, keywords, and other syntactic constructs for which they have no usable level of support. In particular, user agents must not selectively ignore unsupported component values and honor supported values in a single multi-value property declaration: if any value is considered invalid (as unsupported values must be), CSS requires that the entire declaration be ignored.

Experimental implementations

To avoid clashes with future CSS features, the CSS2.1 specification reserves a prefixed syntax for proprietary and experimental extensions to CSS.

Prior to a specification reaching the Candidate Recommendation stage in the W3C process, all implementations of a CSS feature are considered experimental. The CSS Working Group recommends that implementations use a vendor-prefixed syntax for such features, including those in W3C Working Drafts. This avoids incompatibilities with future changes in the draft.

Non-experimental implementations

Once a specification reaches the Candidate Recommendation stage, non-experimental implementations are possible, and implementors should release an unprefixed implementation of any CR-level feature they can demonstrate to be correctly implemented according to spec.

To establish and maintain the interoperability of CSS across implementations, the CSS Working Group requests that non-experimental CSS renderers submit an implementation report (and, if necessary, the testcases used for that implementation report) to the W3C before releasing an unprefixed implementation of any CSS features. Testcases submitted to W3C are subject to review and correction by the CSS Working Group.

Further information on submitting testcases and implementation reports can be found from on the CSS Working Group’s website at http://www.w3.org/Style/CSS/Test/. Questions should be directed to the public-css-testsuite@w3.org mailing list.


Normative References

Bert Bos; et al. Cascading Style Sheets Level 2 Revision 1 (CSS 2.1) Specification. 7 June 2011. W3C Recommendation. URL: http://www.w3.org/TR/2011/REC-CSS2-20110607
Elika J. Etemad; Koji Ishii. CSS Writing Modes Module Level 3. 15 November 2012. W3C Working Draft. (Work in progress.) URL: http://www.w3.org/TR/2012/WD-css3-writing-modes-20121115/
Håkon Wium Lie. CSS3 module: Multi-column layout. CR. URL: http://www.w3.org/TR/css3-multicol/
Gary Kacmarcik; et al. Document Object Model (DOM) Level 3 Events Specification. WD. URL: http://www.w3.org/TR/DOM-Level-3-Events/
Tantek Çelik; et al. Selectors Level 3. 29 September 2011. W3C Recommendation. URL: http://www.w3.org/TR/2011/REC-css3-selectors-20110929/
Rossen Atanassov; Elika Etemad. CSS Fragmentation Module Level 3. WD. URL: http://www.w3.org/TR/css3-break/
Simon Pieters; Glenn Adams. CSS Object Model (CSSOM). WD. URL: http://www.w3.org/TR/cssom/
Anne van Kesteren; et al. W3C DOM4. LCWD. URL: http://www.w3.org/TR/dom/
S. Bradner. Key words for use in RFCs to Indicate Requirement Levels. Best Current Practice. URL: http://www.ietf.org/rfc/rfc2119.txt

Informative References

Tab Atkins Jr.; Elika J. Etemad; Alex Mogilevsky. CSS Flexible Box Layout Module. 18 September 2012. W3C Candidate Recommendation. (Work in progress.) URL: http://www.w3.org/TR/2012/CR-css3-flexbox-20120918/
Koji Ishii. CSS Line Grid Module. Proposal for a CSS module. (Retrieved 26 October 2011) URL: http://dev.w3.org/csswg/css-line-grid/
Alan Stearns. CSS Pagination Templates Module Level 3. Proposal for a CSS module. (Retrieved 4 April 2012) URL: http://dev.w3.org/csswg/css3-page-template/
Melinda Grant; et al. CSS Paged Media Module Level 3. 14 March 2013. W3C Working Draft. (Work in progress.) URL: http://www.w3.org/TR/2013/WD-css3-page-20130314/
Tantek Çelik. CSS3 Basic User Interface Module. WD. URL: http://www.w3.org/TR/css3-ui/
Vincent Hardy; Rossen Atanassov; Alan Stearns. CSS Exclusions Module Level 1. WD. URL: http://www.w3.org/TR/css3-exclusions/
Simon Pieters; Glenn Adams. CSSOM View Module. WD. URL: http://www.w3.org/TR/cssom-view/
Dave Raggett; Arnaud Le Hors; Ian Jacobs. HTML 4.0 Recommendation. REC. URL: http://www.w3.org/TR/html40


Property Index

NameValueInitialApplies toInh.%agesMediaComputed value
flow-intonone | <ident> [element|content]?noneAll elements, but not pseudo-elements such as ::first-line, ::first-letter, ::before or ::after.novisualas specified
flow-from<ident> | none | inheritnoneNon-replaced block containers. This might be expanded in future versions of the specification to allow other types of containers to receive flow content.novisualas specified
break-beforeauto | always | avoid | left | right | page | column | region | avoid-page | avoid-column | avoid-regionautoblock-level elementsnovisualas specified
break-afterauto | always | avoid | left | right | page | column | region | avoid-page | avoid-column | avoid-regionautoblock-level elementsnovisualas specified
break-insideauto | avoid | avoid-page | avoid-column | avoid-regionautoblock-level elementsnovisualas specified
region-fragmentauto | breakautoCSS Regionsnovisualas specified