This document summarizes several plant experiments conducted between 2013-2014. The 2013 Wilting Point Study found that different plant species wilted and died at varying times without water, from 7 to 33 days until wilting and 4 to 80+ days until death. The 2013 Growth Thresholds Study established guidelines for plant growth thresholds by testing 559 plants of 17 species under different light, fertilizer, and moisture conditions. The 2014 Low Light Study validated that plants can survive in very low light and questioned whether the minimum light threshold of 10 DLI was too high for some high light plants.

1.
Plant
Experiments
Summary

2. 2013
Wilting
Point
Study
Conclusions:
Plants
wilted
without
water
after
between
7
days
(Basil)
and
33
days
(Ficus).
The
time
between
wilting
and
death
ranged
from
80+
days
(Aloe)
to
4
days
(Anthurium
and
Rosa).
The
fastest
death
of
a
plant
without
water
was
Basil
(16
days).
The
average
length
of
days
until
a
plant
died
of
lack
of
water
was
24
days
(Anthurium,
Hydrangea,
Rosa,
Spathiphyllum,
Solanum).
Number
of
Plants:
18
Number
of
Species:
9
Number
of
Sensors:
18
FlowerPower
Study
Start
Date:
September,
2013
Study
End
Date:
Mid-­‐September,
2013
Weeks
Total:
2
weeks
Location:
Wageningen
UR,
Bleiswijk,
Netherlands
Purpose:
Determine
the
amount
of
time
plants
of
a
certain
species
can
survive
without
water.
Procedure:
-­‐ Water
plants
thoroughly
at
the
start
of
the
study.
-­‐ Do
not
water
again.
-­‐ Observe
the
decay
and
death
of
the
plant
over
time.
Plant
Distribution:
Setup:
S:PROJETSCONNECTED_DEVICESHawaii19_PlantExperimentWageningen2013
Plant
Name Indoor
Wilting
GROUPS
PER
STUDY
Aloe 2
Anthurium 2
Asplenium
Basil 2
Begonia
Chrysanthemums
Cyclamen
Dracaena
Echeveria
Ficus 2
Hydrangea
Macrophylla 2
Kalanchoe
Phalaenopsis 2
Rosa 2
Solanum
Lycopersicum 2
Spathiphyllum 2
Yucca
Total
Plants 17

3.
2013
Growth
Thresholds
Study
Conclusions:
This
study
was
used
to
determine
guidelines
for
the
growth
thresholds
of
all
plants
in
the
PlantDB.
The
plants
in
this
study
were
considered
‘Model’
or
‘Ambassador’
plants
whose
thresholds
served
as
templates
for
all
the
other
plants.
This
study
served
to
set
the
thresholds
in
fertilizer,
light,
and
soil
moisture
for
all
the
plants
in
the
study.
In
combination
with
the
theoretical
work
performed
by
Filip
van
Noort
of
the
University
of
Wageningen
Greenhouse
Horticulture
Research
Institute
in
Bleiswijk/Zoetermeer
in
the
Netherlands,
this
study
was
used
to
extrapolate
thresholds
towards
all
the
plants
in
the
PlantDB.
This
study
also
served
as
an
empirical
validation
of
the
theoretical
work
performed
by
Dr.
van
Noort.
Number
of
Plants:
559
Number
of
Species:
17
Number
of
Sensors:
150
FlowerPower
Study
Start
Date:
Mid-­‐September,
2013
Study
End
Date:
Mid-­‐November,
2013
Weeks
Total:
8
weeks
Location:
Wageningen
UR,
Bleiswijk,
Netherlands
Purpose:
Determine
the
growth
and
death
thresholds
(for
light,
temperature,
fertilizer
and
moisture)
for
different
types
of
plants
so
as
to
extrapolate
over
as
much
of
the
PlantDB
as
possible.
Procedure:
-­‐ Three
(3)
environments
were
used,
differing
only
in
their
light
levels
(humidity
[80%]
and
temperature
[20
°C
+-­‐
3
°C]
were
controlled
for)
o Indoor
standard
light
level
in
the
North
of
Western
Europe
(0,6
DLI)
o Filtered
light
(4
DLI)
o High
light
(7
DLI)
-­‐ Plants
were
placed
in
conditions
of
soil
humidity,
fertilizer,
and
light
levels
that
were
expected
to
damage
and
eventually
kill
the
plant.
Plant
Distribution:
Setup:
Medium
Light
High
Light
Indoor
S:PROJETSCONNECTED_DEVICESHawaii19_PlantExperimentWageningen2013
Plant
Name Indoor
Light
Medium-­‐Light
Greenhouse
Light
High-­‐Light
Greenhouse
Light
Indoor
Water
Medium-­‐Light
Greenhouse
Water
High-­‐Light
Greenhouse
Water
Medium-­‐Light
Greenhouse
Fertilizer
High-­‐Light
Greenhouse
Fertilizer
Total
Groups Total
Plants
Has
One
Plant
in
All
3
Light
Environments
Aloe 5 15 15 12 4 47 x
Anthurium 5 5 15 3 25 x
Asplenium 5 5 15 12 4 37 x
Basil 5 1 5 NO
-­‐
Basil
Begonia 5 5 15 12 4 37 x
Chrysanthemums 5 5 15 12 4 37 x
Cyclamen 5 5 15 12 4 37 x
Dracaena 5 5 15 12 4 37 x
Echeveria 11 1 11 NO
-­‐
Echeveria
Ficus 5 5 15 12 4 37 x
Hydrangea
Macrophylla 5 5 15 12 4 37 x
Kalanchoe 5 5 15 12 4 37 x
Phalaenopsis 5 15 12 3 32 x
Rosa 5 5 15 12 4 37 x
Solanum
Lycopersicum 5 15 12 3 32 NO
-­‐
Tomato
Spathiphyllum 5 5 15 12 4 37 x
Yucca 5 5 15 12 4 37 x
Total
Plants 55 35 50 60 105 75 108 71

4. 2014
Low
Light
Study
Conclusions:
The
plants
we
tested
survived
remarkably
well
in
even
very
low
light
environments.
This
validated
the
use
of
‘Seasonality’
in
the
design
of
the
alert
system
of
PlantDr.
Also
found
was
that
4
DLI
was
sufficient
to
promote
some
growth
in
Olea
and
Citrus,
asking
the
question
of
whether
the
DLI_MIN
of
10
was
too
high
for
high
light
plants.
As
of
the
writing
of
this
report,
plants
with
the
highest
light
needs
still
have
a
DLI_MIN
of
10.
Number
of
Plants:
132
Number
of
Species:
6
Number
of
Sensors:
60
FlowerPower
Study
Start
Date:
Beginning
of
July,
2014
Study
End
Date:
Mid-­‐December,
2014
Weeks
Total:
24
weeks
Location:
Wageningen
UR,
Bleiswijk,
Netherlands
Purpose:
Determine
the
ability
of
low
light
plants
(chosen
by
popularity
and
expected
resistance
to
low
light)
to
survive
in
different
low
light
environments.
Procedure:
-­‐ Three
(3)
environments
were
used,
differing
only
in
their
light
levels
(humidity
[80%]
and
temperature
[20
°C
+-­‐
3
°C]
were
controlled
for)
o Almost
complete
darkness
(0,1
DLI)
o Indoor
standard
light
level
in
the
North
of
Western
Europe
(0,6
DLI)
o Filtered
light
(4
DLI)
-­‐ Water
and
fertilizer
were
also
controlled
for,
ensuring
that
the
only
variable
at
play
was
the
light
level.
Plant
Distribution:
Setup:
Filtered
Light
Low
Light
Very
Low
Light
Plant
Name Very
Low
Light
(0,1
DLI) Indoor
Light
(0,6
DLI) Filtered
Light
(4
DLI) Total
Plants
by
Species
Ficus
benjamina 10 10 10 30
Yucca
gloriosa 10 10 10 30
Zamioculcas
zamiifolia 10 10 10 30
Howea
forsteriana 10 10 10 30
Citrus
limon 0 0 6 6
Olea
europea 0 0 6 6
Total
Plants
by
Environment 40 40 52 132

5. 2015
Automatic
Watering
DV2
Prototype
Study
(Eco
Mode
Trials)
Conclusions:
This
study
showed
that
in
real
use
cases
the
Kauai
(FP2)
and
FlowerPower
sensors
(FP1)
did
not
have
the
same
VWC
readings
in
the
same
medium.
This
study
was
also
the
first
to
show
that
while
the
pump
was
working,
there
was
not
necessarily
a
proportionate
increase
in
%VWC
seen
by
the
FP2
sensor,
resulting
in
overwatering
(loss
of
water
through
the
drainage
system
and
into
the
‘soucoupe’).
The
only
successful
solution
attempted
in
this
study
to
get
a
better
‘reading’
of
the
waterings
was
to
move
the
sensor.
This
study
highlighted
problems
with
the
battery
consumption
by
the
firmware,
which
was
very
high
(dead
battery
in
one
month).
Results
of
the
Eco
Mode
trials
showed
that
simply
reducing
the
VWC_MIN
thresholds
did
not
equate
to
using
less
water
(Eco_2
used
more
water
in
Tomato
and
Basil
than
Eco_1,
with
better
health
indicators
for
Eco_1).
The
primary
hypothesis
for
this
is
that
it’s
better
to
have
a
slightly
saturated
medium
to
provide
a
binding
agent
for
incoming
water.
A
secondary
hypothesis
rests
on
the
energy
required
by
the
plant
to
draw
water
from
a
dry
medium:
the
drier
the
medium
the
more
energy
used
by
the
plant
to
draw
water
up
from
the
roots,
resulting
in
lower
plant
health
but
also
a
smaller
root
structure
with
less
capacity
to
draw
over
time
–
with
less
of
a
root
structure
and
water
drawing
power,
the
medium
is
less
able
to
hold
water.
Later
studies
(especially
the
Winter
Study
2015
and
Clement
Terrier’s
1er
Canal
studies)
would
show
that
compacting
the
soil
and
taking
into
account
the
difference
in
soil
features
between
a
sodded
plant
and
potting
soil
(when
transplanting
a
plant
into
the
pot)
can
also
account
for
water
loss
and
poor
VWC%
visibility
from
the
FP2
sensor.
Number
of
Plants:
36
Number
of
Species:
3
Number
of
Sensors:
36
Kauai
&
48
FlowerPower
Study
Start
Date:
Beginning
of
July,
2015
Study
End
Date:
Mid-­‐August,
2015
Weeks
Total:
8
weeks
Location:
Wageningen
UR,
Bleiswijk,
Netherlands
Purpose:
Test
the
Firmware
and
Hardware
of
the
Parrot
Pot
DV2
sensor
and
pump.
Procedure:
-­‐ All
plants
of
the
same
species
are
from
the
same
batch
of
seedlings
-­‐ Protocols
are
applied
from
the
very
beginning
of
the
Study
Plant
Distribution:
Setup:
Table Plant Protocol VWC_MIN VWC_CMD
diff VWC_CMD #
of
Plants
Normal_1 17,5 5 22,5 4
Eco_1 15 5 20 4
Eco_2 12,5 5 17,5 4
Normal_1 17,5 5 22,5 4
Eco_1 12,5 5 17,5 4
Eco_2 12,5 5 22,5 4
Normal_1 22,5 5 27,5 4
Eco_1 15 5 20 4
Eco_2 10 5 15 4
3 Tomate
1 Basil
2 Hibiscus

6. 2015
Automatic
Watering
DV2
Prototype
Study
#2
(Eco
Mode
Trials)
Conclusions:
Using
the
same
setups
as
the
Prototype
Study
#1,
this
experiment
confirmed
the
differences
when
reading
%VWC
in
the
same
environment
between
FP1
and
FP2
sensors.
Improvements
to
the
FW
were
confirmed
re:
battery
consumption.
It
was
understood
during
this
study
that
the
%VWC
being
used
by
the
Kauai
after
calibration
was
in
fact
5%VWC
lower
than
what
was
aimed
for
when
setting
the
thresholds
–
so
one
unexpected
conclusion
was
an
understanding
of
just
how
low
the
VWC_MIN
of
basil
could
be
(5%
-­‐
10%)
while
still
growing
healthy
plants.
The
Eco
Mode
protocol
used
in
this
study
differed
from
that
used
in
the
previous
one
(Prototype
Study
#1)
in
that
basil
plants
were
all
grown
using
the
Normal
Protocol
for
one
month
and
then
half
the
plants
were
shifted
to
an
Eco
Mode
Protocol
–
the
result
was
a
lower
health
score
for
the
Eco
Plants
with
no
clear
difference
in
water
consumption
(showing
a
failure
for
the
supposed
‘Eco’
Mode).
Number
of
Plants:
24
Number
of
Species:
1
Number
of
Sensors:
28
Kauai
&
79
FlowerPower
Study
Start
Date:
Beginning
of
November,
2015
Study
End
Date:
End
of
December,
2015
Weeks
Total:
8
weeks
Location:
Wageningen
UR,
Bleiswijk,
Netherlands
Purpose:
-­‐ Test
the
Firmware
and
Hardware
of
the
Parrot
Pot
DV2
sensor
and
pump.
Procedure:
-­‐ The
DV2
Prototype
Pots
that
were
still
working
after
the
first
study
conducted
during
the
summer
of
2015
were
re-­‐used
in
a
second
study
-­‐ Basil
in
a
Normal
protocol
was
used
in
all
cases
for
the
first
half
of
the
study
(until
the
end
of
November)
-­‐ At
the
halfway
point,
half
the
Basil
plants
were
converted
to
an
Eco
protocol
o Plants
were
divided
into
the
two
groups
(Normal
and
Eco)
on
the
basis
of
their
general
health
as
given
by
Wageningen
staff
Plant
Distribution:
Setup:
Plant Protocol Number
of
plants
Basil Eco 12
Basil Normal 12
Evap Normal 4

7. 2015
Automatic
Watering
Winter
DV2
Study
Conclusions:
Les
données
des
Kauai
et
des
FlowerPowers
ont
été
remontes
au
Webservice
par
un
outil
de
Bruno
Sautron
qui
tourne
sur
RaspberryPi
et
est
écrit
en
NodeJS.
Les
problèmes
de
remonté
des
données
vues
sur
le
FB
ont
montré
qu’il
y
avait
des
problèmes
au
niveau
FW
sur
les
uploads.
Cette
étude
a
aussi
utilisé
le
système
de
clignotement
des
LEDS
pour
indiquer
que
tout
allait
bien
(LED
vert)
ou
qu’il
y
avait
un
problème
(LED
rouge).
Ce
système
de
clignotement
des
LEDs
est
la
base
du
système
de
Parrot
Pot
sans
Application.
Six
(6)
Pots
ont
été
remplacés
:
5
pour
problème
de
pompe
KO,
1
pour
problème
de
capteur
KO.
Parmi
les
problèmes
de
pompes
KO,
c’est
la
connexion
sur
la
pompe
qui
est
le
point
faible,
susceptible
de
se
casser
par
oxydation.
Cette
étude
a
encore
confirmé
le
problème
de
cohérence
entre
les
capteurs
FP1
et
FP2
–
ils
ne
voient
pas
le
même
%VWC
dans
le
même
medium.
Il
a
aussi
reconfirmé
le
problème
ou
les
FP2
ne
voient
pas
l’eau
pompé
du
réservoir
&
l’eau
passe
trop
rapidement
à
travers
le
medium
et
dans
la
soucoupe
(elle
n’est
pas
diffue
dans
le
medium,
mais
fuit
rapidement).
Number
of
Plants:
56
Number
of
Species:
6
Number
of
Sensors:
56
Kauai
&
76
FlowerPower
Study
Start
Date:
Beginning
of
November,
2015
Study
End
Date:
End
of
December,
2015
Weeks
Total:
8
weeks
Location:
Wageningen
UR,
Bleiswijk,
Netherlands
Purpose:
-­‐ Test
the
Firmware
and
Hardware
of
the
Parrot
Pot
DV2.
-­‐ Test
the
Flower
Power
2
iOS
Application
and
the
Flower
Power
2
Webservice.
-­‐ Test
the
Flower
Bridge
Developer
tool.
Procedure:
-­‐ All
plants
of
the
same
species
are
from
the
same
batch
of
seedlings
-­‐ Protocols
are
applied
from
the
very
beginning
of
the
Study
Plant
Distribution:
Setup:
Table Plant Protocol VWC_MIN VWC_CMD
diff VWC_CMD #
of
Plants
Normal_1 17,5 5 22,5 4
Normal_2 17,5 10 27,5 4
Eco_1 10 5 15 4
Eco_1bis 10 10 20 4
Eco_2 6 5 11 4
Eco_2bis 6 10 16 4
Normal_1 22,5 5 27,5 4
Normal_2 22,5 10 32,5 4
Eco_1 15 5 20 4
Eco_2 8 5 13 4
Lemon Normal 20 5 25 4
Ficus Normal 17,5 5 22,5 4
Olive Normal 15 10 25 4
Spathiphyllum Normal 17,5 5 22,5 4
3
Basil
Tomato
1
2

8.
2015
HotCold
DV2
Temperature
Study
Conclusions:
Number
of
Plants:
20
Number
of
Species:
5
Number
of
Sensors:
20
Kauai
&
40
FlowerPower
Study
Start
Date:
Mid
December,
2015
Study
End
Date:
Mid
January,
2015
Weeks
Total:
4
weeks
Location:
Wageningen
UR,
Bleiswijk,
Netherlands
Purpose:
-­‐ Test
the
Firmware
and
Hardware
of
the
Parrot
Pot
DV2
under
extreme
temperature
conditions.
-­‐ Test
the
Flower
Power
2
iOS
Application
and
the
Flower
Power
2
Webservice.
-­‐ Test
the
Flower
Bridge
Developer
tool.
Procedure:
-­‐ Two
environments
are
used
o One
‘Hot’
environment
with
an
ambient
temperature
of
35
degrees
Celsius
o One
‘Cold’
environment
with
an
ambient
temperature
of
10
degrees
Celsius
-­‐ Plants
are
all
set
to
their
‘Normal’
protocol
as
defined
by
the
Plant
Distribution:
Setup:
Cold
Environment
Hot
Environment
Environment Plant VWC_MIN VWC_DIFF VWC_CMD Number of Plants
Hot Aloe 15 5 20 3
Hot Ficus 17,5 5 22,5 4
Hot Yucca 17,5 5 22,5 3
Cold Abies 17,5 5 22,5 5
Cold Buxus 17,5 5 22,5 5