diff --git a/examples/Analysis/velocity_autocorrelations.jl b/examples/Analysis/velocity_autocorrelations.jl
index 77ba5e3..b9ac103 100644
--- a/examples/Analysis/velocity_autocorrelations.jl
+++ b/examples/Analysis/velocity_autocorrelations.jl
@@ -10,19 +10,20 @@ space = ContinuousSpace((L,L,L))
 
 model = StandardABM(Microbe{3}, space, Δt; container=Vector)
 n = 200
-for i in 1:n
-    add_agent!(model; turn_rate, motility=RunTumble(speed=[U]))
-    add_agent!(model; turn_rate, motility=RunReverse(speed_forward=[U]))
-    add_agent!(model; turn_rate, motility=RunReverseFlick(speed_forward=[U]))
+for Motility in (RunTumble, RunReverse, RunReverseFlick), i in 1:n
+    add_agent!(model; turn_rate, motility=Motility(speed=[U]))
 end
+ids_runtumble = 1:n
+ids_runreverse = (1:n) .+ n
+ids_runrevflick = (1:n) .+ 2n
 
 nsteps = round(Int, 100τ_run / Δt)
 adata = [:vel]
 adf, = run!(model, nsteps; adata)
 
-adf_runtumble = filter(:id => id -> model[id].motility isa RunTumble, adf; view=true)
-adf_runrev = filter(:id => id -> model[id].motility isa RunReverse, adf; view=true)
-adf_runrevflick = filter(:id => id -> model[id].motility isa RunReverseFlick, adf; view=true)
+adf_runtumble = filter(:id => id -> id in ids_runtumble, adf; view=true)
+adf_runrev = filter(:id => id -> id in ids_runreverse, adf; view=true)
+adf_runrevflick = filter(:id => id -> id in ids_runrevflick, adf; view=true)
 adfs = [adf_runtumble, adf_runrev, adf_runrevflick]
 
 t = range(0, (nsteps-1)*Δt; step=Δt)
diff --git a/test/model-creation.jl b/test/model-creation.jl
index 56f4925..00c6a58 100644
--- a/test/model-creation.jl
+++ b/test/model-creation.jl
@@ -27,16 +27,16 @@ using LinearAlgebra: norm
             rng = Xoshiro(123)
             pos = rand(rng, SVector{D})
             vel = random_velocity(rng, D)
-            speed = random_speed(rng, RunTumble())
+            #speed = random_speed(rng, RunTumble())
+            spd = rand(rng, speed(RunTumble()))
             @test model[1] isa Microbe{D}
-            @test model[1].pos == pos
-            @test model[1].vel == vel
-            @test model[1].speed == speed
-            @test model[1].motility isa RunTumble
-            @test model[1].turn_rate == 1.0
-            @test model[1].radius == 0.0
-            @test model[1].rotational_diffusivity == 0.0
-            @test model[1].state == 0.0
+            @test position(model[1]) == pos
+            @test direction(model[1]) == vel
+            @test speed(model[1]) == spd
+            @test turnrate(model[1]) == 1.0
+            @test radius(model[1]) == 0.0
+            @test rotational_diffusivity(model[1]) == 0.0
+            @test state(model[1]) == 0.0
             # add agent with kwproperties
             motility = RunReverse(
                 speed_backward = [24.0],
@@ -44,7 +44,6 @@ using LinearAlgebra: norm
             )
             add_agent!(model; turn_rate = 0.55, motility)
             @test model[2].turn_rate == 0.55
-            @test model[2].motility isa RunReverse
             @test model[2].speed == 24.0
             # add agent with predefined position
             pos = SVector{D}(i/2D for i in 1:D)
@@ -63,21 +62,22 @@ using LinearAlgebra: norm
                 add_agent!(model)
                 m = model[1]
                 @test m isa T{D}
-                @test m.pos == rand(rng, SVector{D})
-                @test m.vel == random_velocity(rng, D)
-                @test m.speed == random_speed(rng, m.motility)
+                @test position(m) == rand(rng, SVector{D})
+                @test direction(m) == random_velocity(rng, D)
+                @test speed(m) == rand(rng, speed(motilepattern(m)))
                 @test issubset(
                     (:id, :pos, :vel, :speed, :motility,
                     :rotational_diffusivity, :radius, :state),
                     fieldnames(T)
                 )
 
+                φ(microbe, model) = turnrate(microbe) * cwbias(microbe, model)
                 if T == BrownBerg
-                    @test turnrate(m, model) == m.turn_rate * exp(-m.gain*m.state)
+                    @test φ(m, model) == m.turn_rate * exp(-m.gain*m.state)
                 elseif T == Brumley
-                    @test turnrate(m, model) == (1+exp(-m.gain*m.state))*m.turn_rate/2
+                    @test φ(m, model) == (1+exp(-m.gain*m.state))*m.turn_rate/2
                 elseif T == Celani
-                    @test turnrate(m, model) == m.turn_rate * (1 - m.gain*m.state)
+                    @test φ(m, model) == m.turn_rate * (1 - m.gain*m.state)
                     # when no concentration field is set, markovian variables are zero
                     @test m.markovian_variables == zeros(3)
 
diff --git a/test/model-stepping.jl b/test/model-stepping.jl
index b7a3400..e82314c 100644
--- a/test/model-stepping.jl
+++ b/test/model-stepping.jl
@@ -11,10 +11,10 @@ using LinearAlgebra: norm
         model = StandardABM(Microbe{D}, space, dt; rng, container)
         pos = extent ./ 2
         vel1 = random_velocity(model)
-        speed1 = random_speed(rng, RunTumble())
+        speed1 = rand(rng, speed(RunTumble()))
         add_agent!(pos, model; vel=vel1, speed=speed1, turn_rate=0)
         vel2 = random_velocity(model)
-        speed2 = random_speed(rng, RunReverse())
+        speed2 = rand(rng, speed(RunReverse()))
         add_agent!(pos, model; vel=vel2, speed=speed2, turn_rate=Inf, motility=RunReverse())
         run!(model, 1) # performs 1 microbe_step!
         # x₁ = x₀ + vΔt
diff --git a/test/motility.jl b/test/motility.jl
index e12afa4..aa74789 100644
--- a/test/motility.jl
+++ b/test/motility.jl
@@ -3,14 +3,14 @@ using Distributions
 using LinearAlgebra: norm
 
 @testset "Motility" begin
-    @test AbstractMotilityOneStep <: AbstractMotility
-    @test AbstractMotilityTwoStep <: AbstractMotility
-    @test ~(AbstractMotilityOneStep <: AbstractMotilityTwoStep)
+    @test MotilityOneStep <: AbstractMotility
+    @test MotilityTwoStep <: AbstractMotility
+    @test !(MotilityOneStep <: MotilityTwoStep)
 
-    @test fieldnames(RunTumble) == (:speed, :polar, :azimuthal)
+    @test fieldnames(MotilityOneStep) == (:speed, :polar, :azimuthal)
     rt = RunTumble()
     # type hierarchy
-    @test rt isa AbstractMotilityOneStep
+    @test rt isa MotilityOneStep
     # default values
     @test rt.speed == (30.0,)
     @test rt.polar == Uniform(-π, π)
@@ -20,24 +20,21 @@ using LinearAlgebra: norm
     @test rt.speed == Normal(5, 0.1)
     @test rt.polar == [-0.1, 0.1]
     @test rt.azimuthal == (π,)
-    # base constructor without keywords
-    rt2 = RunTumble(Normal(5,0.1), [-0.1, 0.1], (π,))
-    @test rt2.speed == rt.speed && rt2.polar == rt.polar && rt2.azimuthal == rt.azimuthal
 
-    @test fieldnames(RunReverse) == (
-        :speed_forward, :polar_forward, :azimuthal_forward,
+    @test fieldnames(MotilityTwoStep) == (
+        :speed, :polar, :azimuthal,
         :speed_backward, :polar_backward, :azimuthal_backward,
         :motile_state
     )
     # default values
     rr = RunReverse()
-    @test rr isa AbstractMotilityTwoStep
-    @test rr.speed_forward == (30.0,)
-    @test rr.polar_forward == (π,)
-    @test rr.azimuthal_forward == Arccos()
-    @test rr.speed_backward == rr.speed_forward
-    @test rr.polar_backward == rr.polar_forward
-    @test rr.azimuthal_backward == rr.azimuthal_forward
+    @test rr isa MotilityTwoStep
+    @test rr.speed == (30.0,)
+    @test rr.polar == (π,)
+    @test rr.azimuthal == Arccos()
+    @test rr.speed_backward == rr.speed
+    @test rr.polar_backward == rr.polar
+    @test rr.azimuthal_backward == rr.azimuthal
     @test rr.motile_state.state == Forward
     # field overload
     @test rr.state == Forward
@@ -45,19 +42,18 @@ using LinearAlgebra: norm
     rr = RunReverse(; azimuthal_backward = (-π/4,0,π/4))
     @test rr.azimuthal_backward == (-π/4,0,π/4)
     # backward distributions follow forward if unspecified
-    rr = RunReverse(; speed_forward = (45,))
-    @test rr.speed_backward == rr.speed_forward == (45,)
+    rr = RunReverse(; speed = (45,))
+    @test rr.speed_backward == rr.speed == (45,)
 
-    @test fieldnames(RunReverseFlick) == fieldnames(RunReverse)
     # default values
     rrf = RunReverseFlick()
-    @test rrf isa AbstractMotilityTwoStep
-    @test rrf.speed_forward == (30.0,)
-    @test rrf.polar_forward == (π,)
-    @test rrf.azimuthal_forward == Arccos()
-    @test rrf.speed_backward == rrf.speed_forward
+    @test rrf isa MotilityTwoStep
+    @test rrf.speed == (30.0,)
+    @test rrf.polar == (π,)
+    @test rrf.azimuthal == Arccos()
+    @test rrf.speed_backward == rrf.speed
     @test rrf.polar_backward == (-π/2, π/2)
-    @test rrf.azimuthal_backward == rrf.azimuthal_forward
+    @test rrf.azimuthal_backward == rrf.azimuthal
     @test rrf.motile_state.state == Forward
     # field overload
     @test rrf.state == Forward
@@ -65,9 +61,9 @@ using LinearAlgebra: norm
     rrf = RunReverseFlick(; azimuthal_backward = (-π/4,0,π/4))
     @test rrf.azimuthal_backward == (-π/4,0,π/4)
     # polar distributions are independent in run reverse flick
-    rrf = RunReverseFlick(; speed_forward=(45,), polar_forward=(3π,))
-    @test rrf.speed_backward == rrf.speed_forward == (45,)
-    @test rrf.polar_forward == (3π,) && rrf.polar_backward == (-π/2,π/2)
+    rrf = RunReverseFlick(; speed=(45,), polar=(3π,))
+    @test rrf.speed_backward == rrf.speed == (45,)
+    @test rrf.polar == (3π,) && rrf.polar_backward == (-π/2,π/2)
 
     @testset "Motile state" begin
         @test instances(TwoState) == (Forward, Backward)
@@ -105,7 +101,7 @@ using LinearAlgebra: norm
         @test u₁ == u₂
 
         model = StandardABM(Microbe{2}, ContinuousSpace((1,1)))
-        rr = RunReverse(speed_forward=[45], speed_backward=[35]) # initialized to Forward
+        rr = RunReverse(speed=[45], speed_backward=[35]) # initialized to Forward
         add_agent!(model; motility = rr)
         @test random_speed(model[1], model) == 45
         switch!(model[1].motility) # now Backward