Styling ggplot2 graphics

Rolf Bänziger, r.banziger@westminster.ac.uk

17/10/2022

Introduction
We know the three key components of a statistical graphic (data, aesthetics and geometries) and how they
fit in the ggplot2 framework. But these three key components are not enough to produce graphics used in
the real world, and ggplot2 does have additional components. In this lab, we explore how to change the
appearance of graphics.

Style a graphic
Let’s investigate if city MPG and highway MPG are correlated. We map highway mileage (hwy) to the
x-axis and city mileage (cty) to the y-axis, and add a scatter plot layer using geom_jitter. The jitter geom
is a shortcut for geom_point(position = "jitter") and adds a small amount of random variation to the
location of each point to avoid overplotting (i.e. multiple points on the exact same location). Cf. [https:
//ggplot2.tidyverse.org/reference/geom_jitter.html]. We also fit a trend line using linear regressions(method
= "lm") with geom_smooth().
Don’t forget to load the package ggplot2 first.

library(ggplot2)

ggplot(mpg) + aes(cty, hwy) + geom_point() +

geom_smooth(method = "lm")

1
 50

40
hwy

30

20

10 15 20 25 30 35
cty

ggplot(mpg) + aes(cty, hwy) + geom_jitter() +

geom_smooth(method = "lm")

2
 50

40
hwy

30

20

10
10 20 30
cty

Let’s freshen up this graphic a bit. We can set any aesthetic to a fixed setting, e.g. make the dots pink. To
do this, we need to pass the aesthetic as a parameter in the geom_* function (instead of the aes() function).
In the next example, we’re colouring the dots a bit bigger and pink, and making the smoothed line a dashed
black line. Visit Aesthetic specifications for a overview of the available values for each aesthetic.

ggplot(mpg) + aes(cty, hwy) +

geom_jitter(colour = "deeppink", size = 2) +
geom_smooth(method = "lm", colour = "black", linetype = "dashed")

3
 50

40
hwy

30

20

10
10 20 30
cty

Refer to R graph gallery or the figure below for a list of all available colours in R.

4
 white aliceblue antiquewhite antiquewhite1 antiquewhite2
antiquewhite3 antiquewhite4 aquamarine aquamarine1 aquamarine2
aquamarine3 aquamarine4 azure azure1 azure2
azure3 azure4 beige bisque bisque1
bisque2 bisque3 bisque4 black blanchedalmond
blue blue1 blue2 blue3 blue4
blueviolet brown brown1 brown2 brown3
brown4 burlywood burlywood1 burlywood2 burlywood3
burlywood4 cadetblue cadetblue1 cadetblue2 cadetblue3
cadetblue4 chartreuse chartreuse1 chartreuse2 chartreuse3
chartreuse4 chocolate chocolate1 chocolate2 chocolate3
chocolate4 coral coral1 coral2 coral3
coral4 cornflowerblue cornsilk cornsilk1 cornsilk2
cornsilk3 cornsilk4 cyan cyan1 cyan2
cyan3 cyan4 darkblue darkcyan darkgoldenrod
darkgoldenrod1 darkgoldenrod2 darkgoldenrod3 darkgoldenrod4 darkgray
darkgreen darkgrey darkkhaki darkmagenta darkolivegreen
darkolivegreen1 darkolivegreen2 darkolivegreen3 darkolivegreen4 darkorange
darkorange1 darkorange2 darkorange3 darkorange4 darkorchid
darkorchid1 darkorchid2 darkorchid3 darkorchid4 darkred
darksalmon darkseagreen darkseagreen1 darkseagreen2 darkseagreen3
darkseagreen4 darkslateblue darkslategray darkslategray1 darkslategray2
darkslategray3 darkslategray4 darkslategrey darkturquoise darkviolet
deeppink deeppink1 deeppink2 deeppink3 deeppink4
deepskyblue deepskyblue1 deepskyblue2 deepskyblue3 deepskyblue4

The x- and y-axis labels aren’t very useful in the graphic above. To specify labels, use the labs() function.
Any parameters of the label() function are optional.

ggplot(mpg) + aes(cty, hwy) +

geom_jitter(colour = "deeppink", size = 2) +
geom_smooth(method = "lm", colour = "black", linetype = "dashed") +
labs(x = "City mileage", y = "Highway MPG",
title = "Highway by city MPG",
subtitle = "Are highway and city consumption correlated?",
caption = "I \U2665 ggplot2")

5
 Highway by city MPG
Are highway and city consumption correlated?
50

40
Highway MPG

30

20

10
10 20 30
City mileage
I ♥ ggplot2

6
Legends
ggplot2 automatically creates legends when needed. In the next example, we create a graphic where we
encode four variables: engine displacement and highway mileage in the position (x- and y-axes), city mileage
as colour and fuel type as shape.
Because city mileage is a continuous scale, ggplot creates a colour bar legend. Fuel type is discrete, therefore
ggplot2 creates a standard legend.

ggplot(mpg) + aes(displ, hwy, colour = cty, shape = fl) + geom_jitter()

cty
40 35
30
25
20
15
30 10
hwy

fl
c
d
20
e
p
r

2 3 4 5 6 7
displ

Let’s use labs() to change the names of all four variables.

ggplot(mpg) + aes(displ, hwy, colour = cty, shape = fl) + geom_jitter(aes()) +

labs(x = "Engine displacement", y = "Highway MPG",
colour = "City MPG", shape = "Fuel type")

7
 City MPG
40 35
30
25
20
15
Highway MPG

30 10

Fuel type
c
d
20 e
p
r

2 3 4 5 6 7
Engine displacement

Finally, let’s style the legend. The function theme() is used to control the appearance of any element of the
graphic.

ggplot(mpg) + aes(displ, hwy, colour = cty, shape = fl) +

geom_jitter(size = 3, alpha = 0.5) +
labs(x = "Engine displacement", y = "Highway MPG",
colour = "City MPG", shape = "Fuel type") +
theme(legend.position = c(0.9, 0.5),
legend.background = element_rect(fill="lightgrey",
size=0.5),
legend.title = element_text(colour="grey10", size=10,
face="bold"))

8
 City MPG
40 35
30
25
20
15
Highway MPG

30 10

Fuel type
c
d
20
e
p
r

10
2 3 4 5 6 7
Engine displacement

9
Scales
The values for fuel type are not descriptive, let’s change them by adding a custom scale. We use the
scale_shape_manual() function to specify specific shapes (shapes in R are numbered, we use shapes 15
to 19) and the labels for each category. We also make the shapes a bit larger and translucent (size = 3,
alpha = 0.5).

ggplot(mpg) + aes(displ, hwy, colour = cty, shape = fl) +

geom_jitter(size = 3, alpha = 0.5) +
labs(x = "Engine displacement", y = "Highway MPG",
colour = "City MPG", shape = "Fuel type") +
scale_shape_manual(values = 15:19,
labels = c("CNG", "Diesel", "E85", "Premium", "Regular"))

Fuel type
40
CNG
Diesel
E85
Premium
Highway MPG

30 Regular

City MPG
35
30
25
20
20
15
10

2 3 4 5 6 7
Engine displacement

Let’s also specify a custom scale for the colour aesthetic. This time, we use a gradient scale, where we specify
the “low” and “high” colour.

ggplot(mpg) + aes(displ, hwy, colour = cty, shape = fl) +

geom_jitter(size = 3, alpha = 0.5) +
labs(x = "Engine displacement", y = "Highway MPG",
colour = "City MPG", shape = "Fuel type") +
scale_shape_manual(values = 15:19,
labels = c("CNG", "Diesel", "E85", "Premium", "Regular"))+
scale_colour_gradient(low = "palegreen", high = "darkred")

10
 Fuel type
40
CNG
Diesel
E85
Premium
Highway MPG

Regular
30

City MPG
35
30
25
20
20
15
10

2 3 4 5 6 7
Engine displacement

ggplot2 supports many other scales, see [https://ggplot2.tidyverse.org/reference/index.html#section-scales]

for an list of available options.

Logarithmic scales

One commonly used scale is the logarithmic scale. Note how we can specify different scales for the x and
y-axes.

ggplot(mpg) + aes(displ, hwy) + geom_point() +

scale_x_log10() + scale_y_log10()

11
 40

30
hwy

20

2 3 5
displ

ggplot(diamonds[sample(nrow(diamonds), 250),]) + aes(carat, price) + geom_point() +

scale_x_log10() + scale_y_log10()

12
 10000

3000
price

1000

0.3 1.0 3.0

carat

If we want to use a different logarithm, e. g. base2, we need to use scale_*_continuous and specify the
transformation:

ggplot(mpg) + aes(displ, hwy) + geom_point() +

scale_x_continuous(trans = "reverse") +
scale_y_continuous(trans = "log2")

13
 32
hwy

16

7 6 5 4 3 2
displ

Annotations
Sometimes we need to add a label to a chart to explain something. ggplot provides the annotate() function
to add an annotation layer.

ggplot(mpg) + aes(displ, hwy, colour = class) +

geom_jitter() +
geom_smooth(method = "lm", formula = y ~ x, colour = "darkgrey",
size = 0.5, fill = "lightgrey") +
annotate(geom = "text", x= 6.5, y = 29,
label = "2 seaters are\nless economical",
colour = "grey30", size = 3)

14
 40

class
2seater

30 2 seaters are compact

less economical midsize
hwy

minivan
pickup
subcompact
20
suv

10

2 3 4 5 6 7
displ

15
# Facets
While ggplot2 supports several aesthetics, encoding to many variables in one graphic will produce a confusing
chart. Sometimes, it is easier to produce several graphics instead. ggplot2 makes it easy to create a series of
charts with faceting. We just need to add one of the facet functions (facet_wrap() or facet_grid()) and
pass the respective variable wrapped in a vars() call.

ggplot(mpg) + aes(displ, hwy, colour = cty) + geom_jitter() +

facet_wrap(vars(fl))

c d e

40

30

20 cty
35
30
hwy

2 3 4 5 6 7 25
p r
20
40 15
10

30

20

2 3 4 5 6 7 2 3 4 5 6 7
displ

But now, the labels for each chart are meaningless again. To define custom labels, we first need to define a
vector with the labels. Then we set the name of each element to the value in the data.
And of course, we can also still add custom scales.

fl.labels <- c("CNG", "Diesel", "E85", "Premium", "Regular", "x")

names(fl.labels) <- c("c", "d", "e", "p", "r", "x")

ggplot(mpg) + aes(displ, hwy, colour = cty) + geom_jitter() +

scale_colour_gradient(low = "palegreen", high = "darkred") +
facet_wrap(vars(fl), labeller = labeller(fl = fl.labels ))

16
 CNG Diesel E85

40

30

20 cty
35
30
hwy

2 3 4 5 6 7 25
Premium Regular
20
40 15
10

30

20

2 3 4 5 6 7 2 3 4 5 6 7
displ

We can also use two variables in a facet grid.

fl.labels <- c("CNG", "Diesel", "E85", "Premium", "Regular")

names(fl.labels) <- c("c", "d", "e", "p", "r")

ggplot(mpg) + aes(displ, hwy, colour = cty) + geom_jitter() +

scale_colour_gradient(low = "palegreen", high = "darkred") +
facet_grid(fl ~ class, labeller = labeller(fl = fl.labels ))

17
 2seater compact midsize minivan pickup subcompact suv
40

CNG
30
20

40

Diesel
30
20 cty
35

40 30
hwy

E85
30 25
20 20
15
40

Premium
10
30
20

40

Regular
30
20

234567 234567 234567 234567 234567 234567 234567

displ

Themes
Themes are a way to change the look of a graphics completely.

ggplot(mpg) + aes(displ, hwy, colour = class) + geom_jitter() +

theme_minimal()

18
 40

class
2seater
compact
30
midsize
hwy

minivan
pickup
subcompact
suv
20

2 3 4 5 6 7
displ

ggplot2 comes with a some predefined themes. More themes are defined in the package ggthemes. You’ll
need to install it first:

install.packages("ggthemes")

And then we need to load it:

library(ggthemes)

ggthemes provides over 100 themes, palettes and scales.

ls("package:ggthemes")

## [1] "bank_slopes" "calc_pal"

## [3] "calc_shape_pal" "canva_pal"
## [5] "canva_palettes" "circlefill_shape_pal"
## [7] "cleveland_shape_pal" "colorblind_pal"
## [9] "economist_pal" "excel_new_pal"
## [11] "excel_pal" "extended_range_breaks"
## [13] "extended_range_breaks_" "few_pal"
## [15] "few_shape_pal" "fivethirtyeight_pal"
## [17] "gdocs_pal" "geom_rangeframe"
## [19] "geom_tufteboxplot" "GeomRangeFrame"
## [21] "GeomTufteboxplot" "ggthemes_data"
## [23] "hc_pal" "palette_pander"

19
## [25] "ptol_pal" "scale_color_calc"
## [27] "scale_color_canva" "scale_color_colorblind"
## [29] "scale_color_continuous_tableau" "scale_color_economist"
## [31] "scale_color_excel" "scale_color_excel_new"
## [33] "scale_color_few" "scale_color_fivethirtyeight"
## [35] "scale_color_gdocs" "scale_color_gradient_tableau"
## [37] "scale_color_gradient2_tableau" "scale_color_hc"
## [39] "scale_color_pander" "scale_color_ptol"
## [41] "scale_color_solarized" "scale_color_stata"
## [43] "scale_color_tableau" "scale_color_wsj"
## [45] "scale_colour_calc" "scale_colour_canva"
## [47] "scale_colour_colorblind" "scale_colour_economist"
## [49] "scale_colour_excel" "scale_colour_excel_new"
## [51] "scale_colour_few" "scale_colour_fivethirtyeight"
## [53] "scale_colour_gdocs" "scale_colour_gradient_tableau"
## [55] "scale_colour_gradient2_tableau" "scale_colour_hc"
## [57] "scale_colour_pander" "scale_colour_ptol"
## [59] "scale_colour_solarized" "scale_colour_stata"
## [61] "scale_colour_tableau" "scale_colour_wsj"
## [63] "scale_fill_calc" "scale_fill_canva"
## [65] "scale_fill_colorblind" "scale_fill_continuous_tableau"
## [67] "scale_fill_economist" "scale_fill_excel"
## [69] "scale_fill_excel_new" "scale_fill_few"
## [71] "scale_fill_fivethirtyeight" "scale_fill_gdocs"
## [73] "scale_fill_gradient_tableau" "scale_fill_gradient2_tableau"
## [75] "scale_fill_hc" "scale_fill_pander"
## [77] "scale_fill_ptol" "scale_fill_solarized"
## [79] "scale_fill_stata" "scale_fill_tableau"
## [81] "scale_fill_wsj" "scale_linetype_stata"
## [83] "scale_shape_calc" "scale_shape_circlefill"
## [85] "scale_shape_cleveland" "scale_shape_few"
## [87] "scale_shape_stata" "scale_shape_tableau"
## [89] "scale_shape_tremmel" "show_linetypes"
## [91] "show_shapes" "smart_digits"
## [93] "smart_digits_format" "solarized_pal"
## [95] "stat_fivenumber" "stata_linetype_pal"
## [97] "stata_pal" "stata_shape_pal"
## [99] "StatFivenumber" "tableau_color_pal"
## [101] "tableau_div_gradient_pal" "tableau_gradient_pal"
## [103] "tableau_seq_gradient_pal" "tableau_shape_pal"
## [105] "theme_base" "theme_calc"
## [107] "theme_clean" "theme_economist"
## [109] "theme_economist_white" "theme_excel"
## [111] "theme_excel_new" "theme_few"
## [113] "theme_fivethirtyeight" "theme_foundation"
## [115] "theme_gdocs" "theme_hc"
## [117] "theme_igray" "theme_map"
## [119] "theme_pander" "theme_par"
## [121] "theme_solarized" "theme_solarized_2"
## [123] "theme_solid" "theme_stata"
## [125] "theme_tufte" "theme_wsj"
## [127] "tremmel_shape_pal" "wsj_pal"

Let’s see a few examples.

20
Economist

base.plot <- ggplot(mpg) + aes(displ, hwy, colour = class) + geom_jitter()

base.plot +
theme_economist() + scale_colour_economist() +
scale_y_continuous(position = "right")

2seater midsize pickup suv

class
compact minivan subcompact

40

30

hwy
20

2 3 4 5 6 7
displ

Solarized light

base.plot + theme_solarized() + scale_colour_solarized("blue")

21
 40

class
2seater
compact
30
midsize
hwy

minivan
pickup
subcompact

20 suv

2 3 4 5 6 7
displ

Excel 2003

base.plot + theme_excel() + scale_colour_excel()

22
 40

class
2seater
compact
30
midsize
hwy

minivan
pickup
subcompact

20 suv

10
2 3 4 5 6 7
displ

Tableau

base.plot + theme_igray() + scale_colour_tableau()

23
 40

class
2seater
compact
30
midsize
hwy

minivan
pickup
subcompact

20 suv

2 3 4 5 6 7
displ

Colour blind scales

base.plot + scale_colour_colorblind()

24
 40

class
2seater
compact
30
midsize
hwy

minivan
pickup
subcompact

20 suv

2 3 4 5 6 7
displ

Tufte

base.plot + geom_rangeframe() + theme_tufte(base_family = "sans")

25
 40

class
2seater
compact
30
midsize
hwy

minivan
pickup
subcompact

20 suv

10
2 3 4 5 6 7
displ

26
Exercises

CO2 uptake
Use the built-in data set CO2 for the following exercises.

1. Create a scatter plot showing the CO2 uptake by ambient CO2 concentrations. Should you use jitter
or not?
2. Add a dotted trend line using linear regression(“lm”) in your favourite colour. Change the dots to sky
blue large triangles.
3. Encode the origin of the plant in the colour of the triangles. Make sure you still have only one trend
line.
4. Label the chart, axes and legend meaningfully.
5. Use two trend lines (one for each origin). Colour plants from Quebec blue and plants from Mississippi
red.
6. Create the chart above for each Plant using facet_wrap().
7. Create the following chart. Tip: You can hide the legend using the theme() function and its
legend.position parameter.

CO2 uptake
nonchilled chilled

40
CO2 uptake rates (umol/m^2 sec)

30

Quebec
20

10

40

Mississippi
30

20

10

250 500 750 1000 250 500 750 1000

ambient CO2 concentrations (mL/L)

Mammals sleep
Use the msleep data set for the following exercises.

1. Investigate if there is a correlation between body weight and brain weight. You might need to use
logarithmic scales.

27
2. Investigate if the food has an influence on body/brain weight.
3. Does the body weight have an influence on the total sleep? What else has an influence on the sleep
behaviour?

28
Solutions

CO2 uptake

1. Create a scatter plot showing the CO2 uptake by ambient CO2 concentrations. Should you use jitter
or not?

ggplot(CO2) + aes(conc, uptake) + geom_jitter()

40

30
uptake

20

10

250 500 750 1000

conc

2. Add a dotted trend line using linear regression(“lm”) in your favourite colour. Change the dots to sky
blue large triangles.

ggplot(CO2) + aes(conc, uptake) +

geom_jitter(shape = 17, size = 3, colour = "skyblue") +
geom_smooth(method = "lm", colour = "deeppink1", linetype = "dotted")

## `geom_smooth()` using formula 'y ~ x'

29
 40

30
uptake

20

10

250 500 750 1000

conc

3. Encode the origin of the plant in the colour of the triangles. Make sure you still have only one trend
line.

ggplot(CO2) + aes(conc, uptake) +

geom_jitter(aes(colour = Type), shape = 17, size = 3) +
geom_smooth(method = "lm", colour = "deeppink1")

## `geom_smooth()` using formula 'y ~ x'

30
 40

30 Type
uptake

Quebec
Mississippi

20

10

250 500 750 1000

conc

4. Label the chart, axes and legend meaningfully.

ggplot(CO2) + aes(conc, uptake) +

geom_jitter(aes(colour = Type), shape = 17, size = 3) +
geom_smooth(method = "lm", colour = "deeppink1") +
labs(title = "CO2 uptake",
colour = "Origin",
x = "ambient CO2 concentrations (mL/L)",
y = "CO2 uptake rates (umol/m^2 sec)")

## `geom_smooth()` using formula 'y ~ x'

31
 CO2 uptake

40
CO2 uptake rates (umol/m^2 sec)

30 Origin
Quebec
Mississippi

20

10

250 500 750 1000

ambient CO2 concentrations (mL/L)

5. Use two trend lines (one for each origin). Colour plants from Quebec blue and plants from Mississippi
red.

ggplot(CO2) + aes(conc, uptake, colour = Type) +

geom_jitter(shape = 17, size = 3) +
geom_smooth(method = "lm") +
scale_colour_manual(values = c("steelblue3", "firebrick2")) +
labs(title = "CO2 uptake",
colour = "Origin",
x = "ambient CO2 concentrations (mL/L)",
y = "CO2 uptake rates (umol/m^2 sec)")

## `geom_smooth()` using formula 'y ~ x'

32
 CO2 uptake

50
CO2 uptake rates (umol/m^2 sec)

40

Origin
30 Quebec
Mississippi

20

10

250 500 750 1000

ambient CO2 concentrations (mL/L)

6. Create the chart above for each Plant using facet_wrap().

ggplot(CO2) + aes(conc, uptake, colour = Type) +

geom_jitter(shape = 17, size = 3) +
geom_smooth(method = "lm") +
scale_colour_manual(values = c("steelblue3", "firebrick2")) +
labs(title = "CO2 uptake",
colour = "Origin",
x = "ambient CO2 concentrations (mL/L)",
y = "CO2 uptake rates (umol/m^2 sec)") +
facet_wrap(vars(Plant))

## `geom_smooth()` using formula 'y ~ x'

33
 CO2 uptake
Qn1 Qn2 Qn3 Qc1

60

40
CO2 uptake rates (umol/m^2 sec)

20

Qc3 Qc2 Mn3 Mn2

60 Origin
40 Quebec

20 Mississippi

Mn1 Mc2 Mc3 Mc1

60

40

20

250 500 750 1000 250 500 750 1000 250 500 750 1000 250 500 750 1000
ambient CO2 concentrations (mL/L)

7. Create the following chart. Tip: You can hide the legend using the theme() function and its
legend.position parameter.

ggplot(CO2) + aes(conc, uptake, colour = Type) +

geom_jitter() +
geom_smooth(method = "loess", formula = y ~ x) +
scale_colour_manual(values = c("steelblue3", "firebrick2")) +
labs(title = "CO2 uptake",
colour = "Origin",
x = "ambient CO2 concentrations (mL/L)",
y = "CO2 uptake rates (umol/m^2 sec)") +
facet_grid(Type ~ Treatment) +
theme(legend.position = "none")

34
 CO2 uptake
nonchilled chilled

40
CO2 uptake rates (umol/m^2 sec)

30

Quebec
20

10

40

Mississippi
30

20

10

250 500 750 1000 250 500 750 1000

ambient CO2 concentrations (mL/L)

Mammals sleep

Use the msleep data set for the following exercises.

1. Investigate if there is a correlation between body weight and brain weight. You might need to use
logarithmic scales.

ggplot(msleep) + aes(bodywt, brainwt) + geom_point() +

scale_x_log10() + scale_y_log10() +
geom_smooth(formula = y~x, method = "lm") +
labs(x = "Body weight", y = "Brain weight")

## Warning: Removed 27 rows containing non-finite values (stat_smooth).

## Warning: Removed 27 rows containing missing values (geom_point).

35
 1e+01
Brain weight

1e-01

1e-03

1e-01 1e+01 1e+03

Body weight

2. Investigate if the food has an influence on body/brain weight.

ggplot(msleep) + aes(bodywt, brainwt) +

geom_point(aes(colour = vore)) +
geom_smooth(formula = y~x, method = "lm",
size = 0.5, colour= "lightgrey", fill = "grey80") +
scale_x_log10() + scale_y_log10() +
labs(x = "Body weight", y = "Brain weight")

## Warning: Removed 27 rows containing non-finite values (stat_smooth).

## Warning: Removed 27 rows containing missing values (geom_point).

36
 1e+01

vore
carni
Brain weight

1e-01
herbi
insecti
omni
NA

1e-03

1e-01 1e+01 1e+03

Body weight

3. Does the body weight have an influence on the total sleep? What else has an influence on the sleep
behaviour?

vore.labels <- c("Carnivore", "Herbivore", "Insectivore", "Omnivore")

names(vore.labels) <- c("carni", "herbi", "insecti", "omni")

msleep.noNA <- msleep[!is.na(msleep$vore),]

ggplot(msleep.noNA) + aes(sleep_total, bodywt) +

geom_point(aes(colour = genus)) +
geom_smooth(formula = y~x,
method = "lm") +
scale_y_log10() +
labs(x = "Sleep total", y = "Body weight",
title = "Sleep and body weight by food source") +
facet_wrap(vars(vore), labeller = labeller(vore = vore.labels)) +
theme(legend.position = "none")

37
 Sleep and body weight by food source
Carnivore Herbivore
1e+04

1e+01

1e-02
Body weight

Insectivore Omnivore
1e+04

1e+01

1e-02

5 10 15 20 5 10 15 20
Sleep total

38