diff --git a/3.7/indexing-persistent.md b/3.7/indexing-persistent.md
index 50d34fc4ae..5dcb5e567f 100644
--- a/3.7/indexing-persistent.md
+++ b/3.7/indexing-persistent.md
@@ -184,5 +184,3 @@ the sort order of those which are returned can be wrong (whenever the persistent
 index is consulted).
 
 To fix persistent indexes after a language change, delete and re-create them.
-Skiplist indexes are not affected, because they are not persisted and
-automatically rebuilt on every server start.
diff --git a/3.8/indexing-persistent.md b/3.8/indexing-persistent.md
index 50d34fc4ae..5dcb5e567f 100644
--- a/3.8/indexing-persistent.md
+++ b/3.8/indexing-persistent.md
@@ -184,5 +184,3 @@ the sort order of those which are returned can be wrong (whenever the persistent
 index is consulted).
 
 To fix persistent indexes after a language change, delete and re-create them.
-Skiplist indexes are not affected, because they are not persisted and
-automatically rebuilt on every server start.
diff --git a/3.9/http/indexes-multi-dim.md b/3.9/http/indexes-multi-dim.md
new file mode 100644
index 0000000000..f59448d737
--- /dev/null
+++ b/3.9/http/indexes-multi-dim.md
@@ -0,0 +1,7 @@
+---
+layout: default
+---
+Working with multi-dimensional Indexes
+=======================================
+
+{% docublock post_api_index_zkd %}
diff --git a/3.9/indexing-multi-dim.md b/3.9/indexing-multi-dim.md
new file mode 100644
index 0000000000..cec81311f4
--- /dev/null
+++ b/3.9/indexing-multi-dim.md
@@ -0,0 +1,135 @@
+---
+layout: default
+description: A multi dimensional index allows to efficiently intersect multiple range queries
+---
+# Multi-dimensional indexes
+
+The multi-dimensional index type (also called ZKD) provided by ArangoDB can be
+used to efficiently intersect multiple range queries.
+
+A multi-dimensional index is setup by setting the index type to `"zkd"`.
+The `fields` attribute describes which fields are used as dimensions.
+The value of each dimension has to be a numeric (double) value.
+
+## Querying documents within a 3D box
+
+Assume we have documents in a collection `points` of the form
+
+```json
+{"x": 12.9, "y": -284.0, "z": 0.02}
+```
+
+and we want to query all documents that contained within a box defined by
+`[x0, x1] * [y0, y1] * [z0, z1]`.
+
+To do so one creates a multi-dimensional index on the attributes `x`, `y` and
+`z`, e.g. in _arangosh_:
+
+```js
+db.collection.ensureIndex({
+    type: "zkd",
+    fields: ["x", "y", "z"],
+    fieldValueTypes: "double"
+});
+```
+
+Unlike for other indexes the order of the fields does not matter.
+
+`fieldValueTypes` is required and the only allowed value is `"double"`.
+Future extensions of the index will allow other types.
+
+Now we can use the index in a query:
+
+```js
+FOR p IN points
+    FILTER x0 <= p.x && p.x <= x1
+    FILTER y0 <= p.y && p.y <= y1
+    FILTER z0 <= p.z && p.z <= z1
+    RETURN p
+```
+
+## Possible range queries
+
+Having an index on a set of fields does not require you to specify a full range
+for every field. For each field you can decide if you want to bound
+it from both sides, from one side only (i.e. only an upper or lower bound)
+or not bound it at all.
+
+Furthermore you can use any comparison operator. The index supports `<=` and `>=`
+naturally, `==` will be translated to the bound `[c, c]`. Strict comparison
+is translated to their non-strict counterparts and a post-filter is inserted.
+
+```js
+FOR p IN points
+    FILTER 2 <= p.x && p.x < 9
+    FILTER y0 >= 80
+    FILTER p.z == 4
+    RETURN p
+```
+
+## Example Use Case
+
+If you build a calendar using ArangoDB you could create a collection for each user
+that contains her appointments. The documents would roughly look as follows:
+
+```json
+    {
+        "from": 345365,
+        "to": 678934,
+        "what": "Dentist",
+    }
+```
+
+`from`/`to` are the timestamps when an appointment starts/ends. Having an
+multi-dimensional index on the fields `["from", "to"]` allows you to query
+for all appointments within a given time range efficiently.
+
+### Finding all appointments within a time range
+
+Given a time range `[f, t]` we want to find all appointments `[from, to]` that
+are completely contained in `[f, t]`. Those appointments clearly satisfy the
+condition
+
+```
+f <= from and to <= t
+```
+
+Thus our query would be:
+
+```js
+FOR app IN appointments
+    FILTER f <= app.from
+    FILTER app.to <= t
+    RETURN app
+```
+
+### Finding all appointments that intersect a time range
+
+Given a time range `[f, t]` we want to find all appointments `[from, to]` that
+intersect `[f, t]`. Two intervals `[f, t]` and `[from, to]` intersect if
+and only if
+
+```
+a_2 <= b_1 and a_1 <= b_2
+```
+
+Thus our query would be:
+
+```js
+FOR app IN appointments
+    FILTER f <= app.to
+    FILTER app.from <= t
+    RETURN app
+```
+
+## Limitations
+
+Currently there are a few limitations:
+
+- Using array expansions for attributes is not possible (e.g. `array[*].attr`)
+- The `sparse` property is not supported.
+- You can only index numeric values that are representable as IEEE-754 double.
+- A high number of dimensions (more than 5) can impact the performance considerably.
+- The performance can vary depending on the dataset. Densely packed points can
+  lead to a high number of seeks. This behavior is typical for indexing using
+  space filling curves.
diff --git a/3.9/indexing-persistent.md b/3.9/indexing-persistent.md
index 50d34fc4ae..5dcb5e567f 100644
--- a/3.9/indexing-persistent.md
+++ b/3.9/indexing-persistent.md
@@ -184,5 +184,3 @@ the sort order of those which are returned can be wrong (whenever the persistent
 index is consulted).
 
 To fix persistent indexes after a language change, delete and re-create them.
-Skiplist indexes are not affected, because they are not persisted and
-automatically rebuilt on every server start.
diff --git a/3.9/indexing-which-index.md b/3.9/indexing-which-index.md
index 1f8ee3ae92..02d1877228 100644
--- a/3.9/indexing-which-index.md
+++ b/3.9/indexing-which-index.md
@@ -94,6 +94,24 @@ different usage scenarios:
   result TTL indexes will likely not be used for filtering and sort operations in user-land
   AQL queries.
 
+- **multi-dimensional index** (ZKD): a multi dimensional index allows to
+  efficiently intersect multiple range queries. Typical use cases are querying
+  intervals that intersect a given point or interval. For example, if intervals
+  are stored in documents like
+
+  ```json
+  { "from": 12, "to": 45 }
+  ```
+
+  then you can create an index over `from, to` utilize it with this query:
+
+  ```js
+  FOR i IN intervals FILTER i.from <= t && t <= i.to RETURN i
+  ```
+
+  Currently only floating-point numbers (doubles) are supported as underlying
+  type for each dimension.
+
 - **Geo index**: the geo index provided by ArangoDB allows searching for documents
   within a radius around a two-dimensional earth coordinate (point), or to
   find documents with are closest to a point. Document coordinates can either 
@@ -110,7 +128,7 @@ different usage scenarios:
   and a SORT or FILTER statement is used in conjunction with the distance
   function.
 
-- **fulltext index**: a fulltext index can be used to index all words contained in 
+- **fulltext index**: a fulltext index can be used to index all words contained in
   a specific attribute of all documents in a collection. Only words with a 
   (specifiable) minimum length are indexed. Word tokenization is done using 
   the word boundary analysis provided by libicu, which is taking into account 
diff --git a/3.9/indexing.md b/3.9/indexing.md
index 2c50f23727..7144e28c0c 100644
--- a/3.9/indexing.md
+++ b/3.9/indexing.md
@@ -16,5 +16,6 @@ There are special sections for
   - [Persistent Indexes](indexing-persistent.html)
   - [TTL Indexes](indexing-ttl.html)
   - [Fulltext Indexes](indexing-fulltext.html)
+  - [Multi-dimensional Indexes](indexing-multi-dim.html)
   - [Geo-spatial Indexes](indexing-geo.html)
   - [Vertex-centric Indexes](indexing-vertex-centric.html)
diff --git a/_data/3.9-http.yml b/_data/3.9-http.yml
index 3d0aa9f272..337487bacf 100644
--- a/_data/3.9-http.yml
+++ b/_data/3.9-http.yml
@@ -84,6 +84,8 @@
       href: indexes-persistent.html
     - text: TTL
       href: indexes-ttl.html
+    - text: Multi-dimensional
+      href: indexes-multi-dim.html
     - text: Geo-Spatial
       href: indexes-geo.html
     - text: Fulltext
diff --git a/_data/3.9-manual.yml b/_data/3.9-manual.yml
index 483201c871..e6b6668a62 100644
--- a/_data/3.9-manual.yml
+++ b/_data/3.9-manual.yml
@@ -293,6 +293,8 @@
           href: indexing-ttl.html
         - text: Fulltext Indexes
           href: indexing-fulltext.html
+        - text: Multi-dimensional Indexes
+          href: indexing-multi-dim.html
         - text: Geo-spatial Indexes
           href: indexing-geo.html
         - text: Vertex Centric Indexes