diff --git a/xenomorph/xenomorph/number.rkt b/xenomorph/xenomorph/number.rkt
index bb9c5ef8..97807d93 100644
--- a/xenomorph/xenomorph/number.rkt
+++ b/xenomorph/xenomorph/number.rkt
@@ -1,5 +1,5 @@
 #lang debug racket/base
-(require "base.rkt" "int.rkt" "list.rkt" racket/class)
+(require "base.rkt" "int.rkt" "list.rkt" racket/class racket/contract)
 (provide (all-defined-out) (all-from-out "int.rkt"))
 
 #|
@@ -8,6 +8,7 @@ https://github.com/mbutterick/restructure/blob/master/src/Number.coffee
 |#
 
 
+(define (x:float? x) (is-a? x x:float%))
 
 (define x:float%
   (class x:number%
@@ -20,10 +21,22 @@ https://github.com/mbutterick/restructure/blob/master/src/Number.coffee
     (define/augment (x:encode val . _)
       (real->floating-point-bytes val @size (eq? @endian 'be)))))
 
-(define (x:float [size 4] #:endian [endian system-endian]
+(define/contract (x:float [size-arg #f]
+                 #:size [size-kwarg 4]
+                 #:endian [endian system-endian]
                  #:pre-encode [pre-proc #f]
                  #:post-decode [post-proc #f]
                  #:base-class [base-class x:float%])
+  (()
+   ((or/c exact-positive-integer? #false)
+    #:size exact-positive-integer?
+    #:endian endian-value?
+    #:pre-encode (or/c (any/c . -> . any/c) #false)
+    #:post-decode (or/c (any/c . -> . any/c) #false)
+    #:base-class (λ (c) (subclass? c x:float?)))
+   . ->* .
+   x:float?)
+  (define size (or size-arg size-kwarg))
   (new (generate-subclass base-class pre-proc post-proc) [size size] [endian endian]))
 
 (define float (x:float 4))
diff --git a/xenomorph/xenomorph/scribblings/xenomorph.scrbl b/xenomorph/xenomorph/scribblings/xenomorph.scrbl
index 34c89c35..df283188 100644
--- a/xenomorph/xenomorph/scribblings/xenomorph.scrbl
+++ b/xenomorph/xenomorph/scribblings/xenomorph.scrbl
@@ -378,6 +378,55 @@ Little-endian versions of the common integer types. The @racket[u] prefix indica
 
 @subsubsection{Floats}
 
+@defclass[x:float% x:number% ()]{
+Base class for floating-point number objects. By convention, all floats are signed. Use @racket[x:float] to conveniently instantiate new floating-point number objects.
+}
+
+@defproc[
+(x:float?
+[x any/c])
+boolean?]{
+Predicate for whether @racket[x] is an object of type @racket[x:float%].
+}
+
+@defproc[
+(x:float
+[size-arg (or/c exact-positive-integer? #false) #false]
+[#:size size-kw exact-positive-integer? 2]
+[#:endian endian endian-value? system-endian]
+[#:pre-encode pre-encode-proc (or/c (any/c . -> . any/c) #false) #false]
+[#:post-decode post-decode-proc (or/c (any/c . -> . any/c) #false) #false]
+[#:base-class base-class (λ (c) (subclass? c x:float%)) x:float%]
+)
+x:int?]{
+Generate an instance of @racket[x:float%] (or a subclass of @racket[x:float%]) with certain optional attributes.
+
+@racket[size-arg] or @racket[size-kw] (whichever is provided, though @racket[size-kw] takes precedence) controls the encoded size.
+
+@racket[endian] controls the byte-ordering convention.
+
+@racket[pre-encode-proc] and @racket[post-decode-proc] control the pre-encoding and post-decodeing procedures, respectively.
+
+@racket[base-class] controls the class used for instantiation of the new object.   
+}
+
+@deftogether[
+(@defthing[float x:float?]
+@defthing[floatbe x:float?]
+@defthing[floatle x:float?])
+]{
+The common 32-bit floating-point types. They differ in byte-ordering convention: @racket[floatbe] uses big endian, @racket[floatle] uses little endian, @racket[float] uses @racket[system-endian].
+}
+
+@deftogether[
+(@defthing[double x:float?]
+@defthing[doublebe x:float?]
+@defthing[doublele x:float?])
+]{
+The common 64-bit floating-point types. They differ in byte-ordering convention: @racket[doublebe] uses big endian, @racket[doublele] uses little endian, @racket[double] uses @racket[system-endian].
+}
+
+
 @subsubsection{Fixed-points}