Modulation of pheromone trail strength with food quality in Pharaoh's ant, Monomorium pharaonis

Abstract

Pheromone trails are self-organized processes, where colony-level behaviour emerges from the activity of many individuals responding to local information. The Pharaoh's ant is an important model species for investigating pheromone trails. Here we show that Pharaoh's ant foragers mark with trail pheromones, using their stinger, on both the outward and return leg of foraging trips. Examination of trail markings showed that 10.5% of returning fed ants simply made marks by dragging their engorged gaster, because stinger marks were absent. After discounting gaster-dragging hair marks we found that fed ants (42.5%) did not mark significantly more frequently than unfed ants (36.0%). However, we found that trail-marking fed ants marked pheromone trails with a significantly greater intensity, as compared to trail-marking unfed ants, if the food source was high quality (1.0 M sucrose). When the food quality was low (0.01 M sucrose) we detected no significant difference in marking intensity between fed and unfed trail-marking ants. Our results show that in Pharaoh's ants individual trail marking occurs at a frequency of ∼40% among fed and unfed foragers, but the frequency of individuals marking with high intensity (continuous marking) is significantly greater when a food source is high quality. This contrasts with another model species, Lasius niger, where trail strength is modulated by an all-or-nothing individual response to food quality. The reason for this fundamental difference in mechanism is that Pharaoh's ant is highly reliant on pheromone trails for environmental orientation, so must produce trails, whereas L. niger is proficient at visual-based orientation.

Full text

Modulation of pheromone trail strength with food quality
in Pharaoh’s ant, Monomorium pharaonis
DUNCAN E. JACKSON*&NICOLASCHA
ˆLINE†‡
*Department of Computer Science, University of Sheffield
yDepartment of Animal and Plant Sciences, University of Sheffield
zLEEC, UMR CNRS 7153, Universite
´Paris 13, 93430 Villetaneuse, France
(Received 22 June 2006; initial acceptance 19 September 2006;
final acceptance 16 November 2006; published online 22 August 2007; MS. number: 9009R)
Pheromone trails are self-organized processes, where colony-level behaviour emerges from the activity of
many individuals responding to local information. The Pharaoh’s ant is an important model species for
investigating pheromone trails. Here we show that Pharaoh’s ant foragers mark with trail pheromones, us-
ing their stinger, on both the outward and return leg of foraging trips. Examination of trail markings
showed that 10.5% of returning fed ants simply made marks by dragging their engorged gaster, because
stinger marks were absent. After discounting gaster-dragging hair marks we found that fed ants (42.5%)
did not mark significantly more frequently than unfed ants (36.0%). However, we found that trail-marking
fed ants marked pheromone trails with a significantly greater intensity, as compared to trail-marking unfed
ants, if the food source was high quality (1.0 M sucrose). When the food quality was low (0.01 M sucrose)
we detected no significant difference in marking intensity between fed and unfed trail-marking ants. Our
results show that in Pharaoh’s ants individual trail marking occurs at a frequency of w40% among fed and
unfed foragers, but the frequency of individuals marking with high intensity (continuous marking) is
significantly greater when a food source is high quality. This contrasts with another model species, Lasius
niger, where trail strength is modulated by an all-or-nothing individual response to food quality. The rea-
son for this fundamental difference in mechanism is that Pharaoh’s ant is highly reliant on pheromone
trails for environmental orientation, so must produce trails, whereas L. niger is proficient at visual-based
orientation.
Ó2007 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
Keywords: communication; Monomorium pharaonis; Pharaoh’s ant; pheromone trail; self-organization
The foraging trails of ants are often presented as a para-
digm of self-organization (e.g. Camazine et al. 2001). The
behavioural feedback system found in foraging trails per-
fectly illustrates the principle of emergence, where the
activities of many agents responding only to local infor-
mation leads to a global adaptive process. Simple mathe-
matical and computational models have been used to
show how adaptive global solutions can arise in such a sys-
tem, where a trail pheromone provides positive feedback.
However, from a biological perspective, models may have
led to an oversimplified view of this process. Increasingly
research is uncovering great sophistication in the use
of pheromone communication throughout ant trail net-
works. For example, most trail-following ants actually use
multiple trail pheromones secreted from one or more
glands (Witte & Maschwitz 2002; Jackson et al. 2006). Fur-
thermore, behavioural specialization has been found in
the response to particular pheromones and in their selec-
tive deposition (Jackson et al. 2006; Hart & Jackson
2006). Understanding how pheromone trails work clearly
demands a deeper investigation of the behaviour of indi-
vidual ants throughout the foraging process.
The Pharaoh’s ant is an important model species for
conducting empirical research into pheromone trails.
Foraging trails produced by Pharaoh’s ant form complex
branching networks that can extend up to 10 m from the
nest (Sudd 1960). The characteristic geometrical structure
of trail networks is exploited as an aid in orientation,
whilst the long-lived nature of the pheromone used in
Correspondence: D. E. Jackson, Department of Computer Science, Uni-
versity of Sheffield, Regent Court, 211 Portobello Street, Sheffield S1
4DP, U.K. (email: duncan@dcs.sheffield.ac.uk).
463
0003e3472/07/$30.00/0 Ó2007 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
AN IM AL B EH A VI OU R, 2007, 74, 463e470
doi:10.1016/j.anbehav.2006.11.027

networks facilitates rapid exploration of the environment
from day-to-day (Jackson et al. 2004, 2006). Pheromone
trails of Pharaoh’s ant also contain a short-lived compo-
nent that lasts for only 25 min (Jeanson et al. 2003),
which contrasts with the long-lived component that lasts
for over 2 days (Jackson et al. 2006). Seven pheromones
known to elicit trail following in Pharaoh’s ant have
been chemically isolated and synthesized. Monomorines,
isolated from the poison glands, have trail following activ-
ity at concentrations as low as 10
8
g/cm, with blends of
monomorines I and III being the most active (Ritter
et al. 1977a). Faranal, found in trace amounts in the Du-
four’s glands (Ritter et al. 1977b), is active at concentra-
tion as low as 10
12
g/cm, although maximal activity
occurs at 10
9
g/cm (Ritter et al. 1977b; Ho
¨lldobler &
Wilson 1990). Trail pheromones are deposited on surfaces
using the extruded stinger. Both the poison gland and the
Dufour’s gland empty through the stinger, but secretions
from either gland can be deposited separately (or simulta-
neously) because both glands possess reservoirs with
sphincters to control release (Billen 1987).
It is frequently assumed that ants only lay pheromone
after discovering food, whilst they are returning to the
nest. This is only true of a returning ant that can navigate
back to the nest using her individual orientation skills,
such as memory of visual landmarks, path integration or
a sun compass (Ho
¨lldobler & Wilson 1990). Once the suc-
cessful forager returns to the nest its pheromone trail can
then guide recruited ants to the food source. However,
many ant species are reliant on pheromone trails for nav-
igation during foraging and are obliged to lay trail pher-
omone to guide them back to the nest. Pharaoh’s ant is
reliant on pheromone trails for orientation in the forag-
ing environment. Foragers on active Pharaoh’s ant trails
are known to deposit trail pheromone both when fed
and unfed, and when walking to or from the nest (Hart
& Jackson 2006). It has been observed that Pharaoh’s
ant produces ‘exploratory’ trails in the absence of food,
but only after the discovery of food are such trails termed
foraging trails, or ‘exploitation’ trails (Fourcassie
´& De-
neubourg 1994). This terminology is confusing and sug-
gests that there are qualitative differences between
exploratory trails and foraging trails, where none has
been shown.
A major problem in understanding pheromone trail
dynamics is determining the contribution individual ants
make to a pheromone trail and how that contribution
leads to the colony-level selection of food sources. Pre-
vious studies have suggested that fed ants contribute most
to trails but these studies have relied on indirect measures
of trail deposition, particularly behavioural observation of
body postures such as gaster curling that might indicate
trail marking (Beckers et al. 1993; Mailleux et al. 2000). We
determined the contribution that fed and unfed Pharaoh’s
ant foragers made to pheromone trails, but used a direct
measure of pheromone deposition. We studied the physi-
cal marks made by ants walking on a smoked glass surface
(Hangartner 1969; Jackson et al. 2004). Using this direct
measure we analysed trail-marking frequency on active
foraging trails and the intensity of marking with variable
food quality.
METHODS
Study Species
Four study colonies of Pharaoh’s ant (Formicidae:
Myrmicinae) each contained 1200e2500 workers, brood
of all stages and multiple queens (12e50). Pharaoh’s ants
are small (ca. 2 mm in length), monomorphic ants and
easily maintained in the laboratory. Colonies were
housed in wooden nestboxes (11 8 cm and 2 cm high)
held within a large plastic foraging box (45 30 cm and
15 cm high) in a climate controlled room (24 2C,
RH ¼30%, 12:12 h light:dark cycle). Colonies were given
fresh water ad libitum in glass tubes sealed with cotton
wool and fed with Tenebrio larvae, sugar syrup, dried liver
and dried egg yolk.
Trail-marking Frequency
To determine the frequency of pheromone trail marking
when walking to and from a feeder, we constrained ants to
produce pheromone trails in a narrow corridor placed
upon smoked glass (Fig. 1). A sheet of toughened glass
(74 28 cm) was held over a wax candle flame so as to
coat it with a fine layer of soot. A glass sheet was then
placed on top of a clear glass tank. A colony accessed
one edge of the smoked glass via a bridge, which led to
a corridor (1 cm internal width) made of two polycarbon-
ate strips (60 4 cm and 0.5 cm). The inner surfaces of the
corridor were coated with Fluon to prevent ants from
climbing them. A syrup feeder was placed at the far end
of the corridor. Observations began when colonies had
successfully established foraging trails, which we opera-
tionally defined as when 20 ants had returned from the
feeder. This was to ensure pheromone trails were not
heavily marked before we began observations, but also
to ensure a continuous trail had been established. In the
early stages of the trail establishment process the trails
of ants are still very weak, or possibly discontinuous,
and trail following behaviour on such unestablished trails
may differ from that on strong, established trails. We
chose the point after which 20 ants had returned from
the feeder because trail traffic was regular at this point, in-
dicating trail establishment.
Strong light sources were placed above and below the
corridor to facilitate observation of marking by individual
ants. Individuals were observed as they walked the full
length of the corridor from the feeder to the bridge or vice
versa. The corridor usually contained multiple ants,
because this was an active trail. Focal ants were randomly
selected as they either left the feeder, or stepped off the
bridge into the corridor, and they were followed until they
either reached the bridge (fed) or feeder (unfed). We
observed each ant continually using a 20magnifying
hand lens, which was moved along the top of the corridor
as the focal ant walked along. We paid close attention to
the smoked glass for any markings made by the focal ant.
Any trail marking was recorded (excluding footprint
marks). We observed 50 ants walking towards the food
source (unfed) and 50 ants returning to the nest from the
ANIMAL BEHAVIOUR, 74,3
464

feeder (fed). This procedure was repeated for four colonies
using a fresh sheet of smoked glass for each colony.
Trail Marking
We examined the isolated trail markings of individual
ants to discriminate between different types of marking
and also to measure the intensity of trail marking. We
modified the apparatus shown in Fig. 1 and made the cor-
ridor 2.5 cm wide. Once colonies had established a foraging
trail (>20 ants returned from the 1.0 M sugar syrup feeder)
we placed a smoked glass microscopy slide (7.5 2.5 cm)
lengthways in the middle of the widened foraging corridor.
When an ant stepped on the slide, the slide was carefully
removed from the corridor using forceps and elevated to
a height of approximately 20 cm. The slide was kept in
a horizontal position to minimize disturbance to the ant.
We removed the slide to ensure only one ant walked on
the slide. We allowed the ant to walk across the slide until
it reached an edge at which point the ant was shaken from
the slide. This procedure was repeated using four colonies
in total, and for each colony we tested 50 unfed ants and
50 fed ants. A fresh slide was used for each ant tested.
The same experimental operator always performed the
slide manipulation so as to eliminate operator bias. Each
slide was suspended above a bright light source and exam-
ined with a 20magnifying lens.
Using Hangartner’s (1969) scheme we classified the
markings made by individual Pharaoh’s ants. Hangartner
(1969) showed that the fire ant, Solenopsis geminata, pro-
duces easily classified trail markings on smoked glass, de-
pending on whether the stinger is extended and the
degree of contact with the substrate. We also measured
the distance marked per unit distance walked upon the
slide. Markings were classified as either trail markings
(continuous streaks and discontinuous dots); hair mark-
ings produced by gaster dragging; or footprints. Combina-
tions of these markings were also observed. Intensity of
trail marking was measured as the proportion of the total
distance walked on the slide when trail markings were
made. To do this we placed transparent plastic over the
back-lit slide and traced the path of the ant onto the plas-
tic with a pen. Some ants made sinuous paths across the
slide, so intensity was normalized to give total trail
marked as a proportion of total distance walked (trail
marking per 1 cm walked).
A further measure of intensity, also first noted by Hangart-
ner (1969), is whether ants mark with continuous lines or
discontinuous dots. These different markings indicate vari-
ability in pressure applied by the stinger to a surface and
this determines the amount of pheromone likely to be
deposited. We quantified the number of fed and unfed
ants that marked continuously or discontinuously on slides.
Trail Marking with Low Quality Food Source
We repeated this trail marking investigation procedure
with a low quality (0.01 M) sugar syrup feeder using two
colonies, each time for 50 fed and 50 unfed ants.
Classification of Trail Markings
We produced images of the different markings made by
Pharaoh’s ant by allowing a single colony to access a sheet
of smoked glass using a bridge. The set-up used was as
shown in Fig. 1 but the corridor was omitted, which al-
lowed ants to forage without constraint. The smoked glass
sheet was removed after 2 h and cleared of ants by shak-
ing. We then produced an image of trail markings on
the glass by scanning it using a Canon Canoscan Lide
Scanner, with backlighting from a Canon slide copier
illuminator.
RESULTS
Trail-marking Frequency
Table 1 shows the trail-marking frequency of fed and
unfed ants from four test colonies. We found no signifi-
cant colony effect (two-way ANOVA: F
3,392
¼0.32,
P¼0.814) on marking frequency, but found a highly
significant effect of fed/unfed status (F
1,392
¼15.81,
Nestbox
Syrup feeder
Corridor
Bridge
Foraging box
Light source
Light source
Glass tank
Smoked glass sheet
Figure 1. Experimental apparatus used for investigation of pheromone trails on smoked glass. Ants accessed the corridor leading to the sugar
syrup feeder, via a bridge from the foraging box. The corridor constrained foraging ants to form a straight trail to the feeder on the smoked
glass surface, whilst strong illumination from both above and below facilitated observation of trail marking. For closer examination of trail mark-
ings a smoked glass slide (7.5 2.5 cm) was placed lengthways in the centre of the corridor and removed with forceps when an ant stepped
onto it. An ant was allowed to walk on the slide until it reached an edge. Fifty fed and 50 unfed ants from each of four colonies were tested.
JACKSON & CHA
ˆLINE: ANT PHEROMONE TRAIL MARKING 465

P<0.001). There was no significant interaction effect be-
tween colony and fed/unfed status (colony*fed/unfed:
F
3,392
¼0.84, P¼0.471). Overall (colonies pooled) 54.5%
of fed ants made trail markings compared to only 35.0%
unfed ants marked.
Trail Markings
Categories of pheromone trail markings made by Phar-
aoh’s ant are shown in Fig. 2. Markings made by foragers
took the form of discontinuous trails (spots: Fig. 2d) con-
tinuous trails (streaks: Fig. 2b, f), footprints (Fig. 2c), hair
marks (Fig. 2e) or combinations of these markings.
When we examined markings made on unconstrained
smoked glass (no corridor) we also found that the regions
between trails (Fig. 2a) were marked with discontinuous
sections of trail and infrequent spots indicating sparing
deposition of trail pheromone.
Markings made in trials by four colonies are presented in
Table 2. The categories we used for classification of mark-
ings were trail markings (continuous and discontinuous
stinger markings or combined stinger plus hair markings);
hair marks alone; and footprints alone (no trail marking).
When hair marks were included as trail markings results
were very similar to those shown in Table 1. We found no
significant colony effect (two-way ANOVA: F
3,392
¼0.19,
P¼0.903) on marking frequency, but found a highly
significant effect of fed/unfed status (F
1,392
¼10.45,
P¼0.001). There was no significant interaction effect be-
tween colony and fed/unfed status (colony*fed/unfed:
F
3,392
¼0.35, P¼0.786). However, when we excluded
hair marking as trail markings we found there was no
significant difference in the trail marking (stinger marks)
frequency of fed and unfed ants (F
1,392
¼1.75, P¼0.186).
We found no significant colony effect (F
3,392
¼0.34,
P¼0.795) and there was no significant interaction effect
between colony and fed/unfed status (colony*fed/unfed:
F
3,392
¼0.29, P¼0.835). Markings with abdomen hairs
(stinger markings absent) were made by 10.5% of all
fed ants (colonies pooled) compared to only 1% of all unfed
ants (colonies pooled), and were a consequence of fed ants
dragging their engorged gaster.
We found a significant difference in intensity of mark-
ing between trail-marking fed and unfed ants, using two
separate measures. We found that fed ants laid trail
(stinger marking) with a slight but significantly greater
intensity (as proportion of distance marked) compared to
unfed trail-laying ants (Student’s ttest: t¼2.06,
P¼0.044). Unfed ants made trail markings for 0.64 cm
(SD ¼0.36) of every 1.0 cm walked, whilst fed ants
marked 0.75 cm per 1.0 cm (SD ¼0.31). In our second
measure of marking intensity we found that trail-marking
fed ants made continuous markings on glass slides signif-
icantly more frequently (72.9%) than trail-marking unfed
(50.0%) ants (chi-square test: c2
1¼8:75, N¼157,
P¼0.003).
Trail Marking with Low Quality Food Source
Table 3 shows results obtained when ants foraged to
a 0.01 M feeder. We found no significant effect of fed/
unfed status (two-way ANOVA: F
1,196
¼0.33, P¼0.564)
on marking frequency. There was no colony effect
(F
1,196
¼0.75, P¼0.387) and no significant interaction
effect between colony and fed/unfed status (colony*fed/
unfed: F
1,196
¼0.08, P¼0.773). The intensity of marking
as measured in marking per unit distance did not differ
significantly between fed (0.68 cm marked per 1 cm,
SD ¼0.41, N¼40) and unfed (0.60 cm marked per 1 cm,
SD ¼0.38, N¼36) ants, with the low quality (0.01 M)
feeder (Student’s ttest: t¼0.879, P¼0.382). In our second
analysis of intensity we found that only 55% of trail-
marking fed ants (22/40) marked continuously, which
was not significantly different to the frequency of contin-
uously marking unfed trail-marking ants (16/36 ¼44.4%;
chi-square test: c2
1¼0:844, N¼76, P¼0.358).
The frequency of trail-marking fed ants that continu-
ously marked when a 1.0 M feeder was present (62/
85 ¼72.9%) was significantly greater than when
a 0.01 M feeder (22/40 ¼55%) was used (chi-square test:
c2
1¼3:962, N¼125, P¼0.046). We found no difference
between the proportions of unfed trail-marking ants that
continuously marked with the two different feeder quali-
ties (c2
1¼0:297, N¼108, P¼0.586). Our results suggested
that w50% of unfed trail-marking ants always marked
trails with a consistent intensity, but that trail-marking
fed ants increased the intensity of their marking as a func-
tion of feeder quality. This increased intensity was evident
in an increased proportionate distance for which trail was
laid and also in an increased proportion of ants marking
continuously, when a 1.0 M feeder was used.
DISCUSSION
Our data confirm that Pharaoh’s ant foragers mark pher-
omone trails whilst walking to food and the nest. Initially,
we found that the frequency of trail marking was higher in
fed ants (54.5%) than unfed ants (35.0%). However, 10.5%
of fed ants were making parallel hair markings (Fig. 2e) by
dragging their engorged gaster. These hair markings were
not accompanied by marking with the stinger, which is
the source of all trail pheromones in Pharaoh’s ant
Table 1. Frequency of pheromone trail marking (þ) by individual
ants foraging within a 60 cm corridor placed on a sheet of smoked
glass
Unfed Fed
Trail marks
þ
Trail marks

Trail marks
þ
Trail marks

Colony 1 15 35 27 23
Colony 2 17 33 29 21
Colony 3 20 30 23 27
Colony 4 18 32 30 20
All colonies 70 (35%) 130 (65%) 109
(54.5%)
91 (45.5%)
We recorded no trail marking () when only footprints were added
to markings already present in the soot. We observed 50 fed ants
walking to the nest and 50 unfed ants walking to the food source
from each of four colonies.
ANIMAL BEHAVIOUR, 74,3
466

(monomorine trail pheromones were chemically unde-
tectable in hair markings, D.E. Jackson, unpublished
data). When hair markings were excluded from the analy-
sis we found fed and unfed ants marked trails with equal
frequency. We found a small, but significant, difference
in the intensity of trail marking, where trail-marking fed
ants laid trail for a greater proportion of the distance
they walked on a slide. However, a more significant mea-
sure of intensity was that 50% of unfed marking ants
laid discontinuous trail markings, but only 17.1% of fed
marking ants made such intermittent markings. Fed and
unfed Pharaoh’s ant workers laid trail with equal probabil-
ity on an established trail to a high quality (1.0 M sucrose)
feeder, but trail-marking fed ants laid trail with a much
greater intensity. When a low quality food source
(0.01 M sucrose) was present fed and unfed ants marked
with the same frequency and the same intensity.
With a 1.0 M sucrose feeder we found that overall 39.3%
of foragers (fed and unfed pooled) made trail markings in-
dicative of pheromone deposition, which closely matches
the proportion of ants (43.0%) found to make frequent
U-turns on trails in the study of Hart & Jackson (2006).
U-turners were shown to be specialized ants with a high
probability (88%) of trail marking. In Lasius niger, a species
where trail-laying behaviour (not physical marking) has
been extensively investigated, trail-laying frequency varies
from 34% to 77% on inbound and outbound legs of forag-
ing trips, but the frequency declines as foraging bouts
Figure 2. Markings formed by a Monomorium pharaonis colony foraging on a large (39.2 27.6 cm) sheet of smoked glass, but without the
corridor shown in Fig. 1. (a) Discontinuous, spotted markings typical of off-trail regions (bar ¼0.25 mm); (b) continuous streaked markings
that form a main trail (bar ¼1.0 mm); (c) area of intense footprint marking only (bar ¼0.20 mm); (d) long section of spotted trail
(bar ¼0.15 mm); (e) parallel markings made by hairs whilst dragging the abdomen (bar ¼0.02 mm); and (f) continuous, unbroken streak
typical of main trail (bar ¼0.25 mm).
JACKSON & CHA
ˆLINE: ANT PHEROMONE TRAIL MARKING 467

progress (Detrain et al. 2001; Mailleux et al. 2005). It seems
unlikely that a similar decline in marking frequency would
be found during Pharaoh’s ant foraging bouts. This is be-
cause pheromone trails are essential for the orientation
of Pharaoh’s ant, when searching for food and transporting
food back to the nest. Because of this reliance on trails
Pharaoh’s ant forms an ‘exploratory trail’ network when-
ever new territory is encountered (Fourcassie
´& Deneu-
bourg 1994). Sections of these exploratory trails become
active ‘exploitation’ trails once food is discovered. How-
ever, this frequently used distinction between ‘explor-
atory’ and ‘exploitation’ trails is an artificial one and
a source of confusion. In every respect all pheromone trails
formed by foragers are simply foraging trails, whether they
currently lead to food sources or not. Pharaoh’s ant is ob-
liged to maintain pheromone trails throughout foraging
persists but L. niger can use visual cues, such as landmarks
and polarised light, to navigate between the nest and food
(Aron et al. 1993). We suggest that the proportion of indi-
viduals engaged in trail marking does not vary in Pharaoh’s
ant because pheromone trails are so essential for
orientation.
We found that trail marking intensity varied signifi-
cantly between fed and unfed Pharaoh’s ant foragers with
a 1.0 M feeder, but there was no difference in intensity of
marking, between fed and unfed ants, when a 0.01 M
feeder was present. In L. niger the probability of trail mark-
ing in response to a food source varies idiosyncratically be-
tween individuals as a function of the food quality
(Mailleux et al. 2005). Individuals will only lay trail pher-
omone if the quality of a food source satisfies, or exceeds,
their personal selection criteria (Mailleux et al. 2000).
Thus if a food source is of poor quality, fewer individuals
will choose to mark and the pheromone trail will be
weak. In L. niger this all-or-nothing response is the key
component of trail strength modulation (Mailleux et al.
2003). Our results with Pharaoh’s ant are similar to the sit-
uation found in the fire ant, Solenopsis geminata, where
a constant proportion of individuals mark with trail pher-
omone, but they can modulate trail strength through vari-
able intensity of deposition. Trail deposition intensity is
modulated by varying the pressure of stinger contact
with the surface (Hangartner 1969). The result is that
either strong, continuous marks or weak, discontinuous
Table 2. Intensity and frequency of trail marking by individual ants on smoked glass slides
Unfed Fed
Footprints
only Trail marks
Trail marked
per cm
Hair marks
only
Footprints
only Trail marks
Trail marked
per cm
Hair marks
only
Colony 1 30 19 0.620.33 1 22 20 0.730.32 8
Colony 2 32 17 0.660.37 1 25 19 0.820.28 6
Colony 3 34 16 0.610.39 0 22 23 0.740.31 5
Colony 4 30 20 0.660.36 0 25 23 0.710.32 2
Total 126 72 d29485d21
Mean
(%)SD
633.82 363.65 0.640.36 11.15 473.46 42.54.12 0.750.31 10.55.00
Food source was a 1.0 M sucrose syrup feeder. Fifty fed (returning) and 50 unfed (outbound) foraging ants, from each of four colonies were
observed (fed ants: total N¼200; unfed ants: total N¼200). Trail markings were those made either by the stinger alone (continuous streaks,
short dashes or dots) or a combination of hair and stinger markings. No trail marking was characterised by footprints alone. Hair markings were
those made by abdomen hairs alone and were a consequence of engorged ants dragging their gaster. For trail-laying ants (stinger marking) we
measured the proportion of the total distance walked in crossing the slide where trail was laid, as a measure of intensity (distance trail laid per
cm walked).
Table 3. Intensity and frequency of trail marking by individual ants on smoked glass slides, where the food source was a 0.01 M sucrose syrup
feeder
Unfed Fed
Footprints
only Trail marks
Trail marked
per cm
Hair marks
only
Footprints
only Trail marks
Trail marked
per cm
Hair marks
only
Colony 1 28 20 0.600.38 2 27 21 0.690.38 2
Colony 2 34 16 0.590.38 0 29 19 0.670.43 2
Total 62 36 d25640d4
Mean
(%)SD
628.5 365.7 0.600.38 22.8 562.8 402.8 0.680.41 40
Fifty fed (returning) and 50 unfed (outbound) foraging ants, from each of two colonies were observed (fed ants: total N¼100; unfed ants:
total N¼100).
ANIMAL BEHAVIOUR, 74,3
468

(spotted) marks are made. The surface area of the stinger
making contact with the substrate determines the volume
of pheromone deposited. In S. geminata the intensity of in-
dividual marking was shown to be modulated with food
quality, food distance and volume consumed. As with
Pharaoh’s ant, foragers of S. geminata are reliant on trails
for orientation. Our results show that trail pheromone
must be continually deposited throughout foraging bouts
in species that rely on trails for orientation. A variation in
trail marking intensity with food quality is found in Phar-
aoh’s ant, where fed ants marked trail with greatest inten-
sity but this intensity was not different to unfed when
a lower quality food source was present. Our results sug-
gest that individual responses to food must be the key to
modulating trail strength in Pharaoh’s ant. This contrasts
with L. niger, where individual choice as to whether they
mark, or not, is most important.
The two different methods of modulating trail phero-
mone concentration have important consequences for the
performance of ant colonies when faced with different
foraging challenges. But modulation of trail concentration
through variation of individual intensity of marking and
all-or-nothing individual choice both achieve the same
end for the colony; a graded response to food quality.
However, for an obligate (or highly reliant) trail-following
ant species a modulation of trail marking intensity is the
only viable option. But how does this affect foraging
performance? Sumpter & Beekman (2003) showed that
Pharaoh’s ant selected the better of two food sources in
14 out of 18 trials and suggested that ‘the amount of pher-
omone that the ants deposit.depends on the strength of
the food sources’. We have confirmed this suggestion and
shown that the mechanism is quite different from that
found in the model species L. niger. Likewise Beckers
et al. (1993) found that L. niger selected the best feeder
in 12 out of 14 trials. So both strategies perform almost
equally when presented with a simple binary choice be-
tween two unequal food sources. However, a challenge
of greater importance from an ecological viewpoint is
the ability to select and concentrate efforts on the best
available resource in a changing, competitive environ-
ment. In L. niger switching to a better resource is either
very slow, or never occurs, when the introduction is after
an initial resource has been selected (Beckers et al. 1990).
Lasius niger usually becomes ‘stuck’ on an initial poor
resource even though a better one is available. In L. niger
visual memory of routes play a major part in foraging so
switching from one source to another may require more
than the decay of a pheromone trail. Sudd (1960) showed
that Pharaoh’s ant is slow to redirect workers when better
food resources are made available, but colonies can usually
do so in about an hour. However, Sudd (1960) was work-
ing in a natural environment with trails of up to 10 m
in length, compared to trails typically of less than 1 m
used in experimental studies of L. niger (Beckers et al.
1990). Sudd (1960) also observed that in Pharaoh’s ant
it is commonplace for a single colony to forage on multi-
ple resources simultaneously, using several pheromone
trails.
A recognized benefit of trail marking in both di-
rections is that the shortest path to food can be selected
without recourse to visual information (Goss et al.
1989). In a binary choice trail set-up, where two differ-
ent length paths to the same food source, the shortest
route can be selected when ants mark in both direc-
tions, because for an initial period the shortest branch
is marked by ants moving in both directions whilst
the longest branch is only marked by ants walking in
one direction during that period. Thus the shortest
branch accumulates pheromone faster and is selected.
The greater the difference in the amount of marking
made by ants walking to the food and those returning
is, then the more random the overall choice between
two different length branches. This mechanism predicts
that L. niger will be poor at selecting the shortest of two
paths, because it only marks trails after food is found,
whereas in fact L. niger performs well. The solution to
this apparent paradox is that L. niger can use additional
information, such as visual cues, and can identify the
most direct route back to the nest (Beckers et al. 1993;
Camazine et al. 2001). Lasius niger foragers remember
the general direction to the nest and those returning
on the longer path can turn around, and then follow
the shorter route, when they see that the route they
are taking does not lead directly to the nest. However,
some individuals also prefer to return to the nest using
the same route they followed to the food (Beckers et al.
1992). It is clear that a foraging mechanism which is re-
liant on pheromones performs just as well as one with
recourse to visual memory.
Pheromone trails enable rapid mass recruitment to food
discoveries (Ho
¨lldobler & Wilson 1990) but they also
impose constraints on the overall foraging strategy of
a species. The characteristics of trail pheromones used,
particularly their decay rate, impose major constraints
on foraging flexibility. For example, if pheromone decay
is slow the ability to switch to better food sources is simi-
larly slow. Ant species vary in their reliance on pheromone
trails and visual cues for orientation (Aron et al. 1993). A
high reliance on pheromone trails requires that trails are
marked constitutively as an essential aid for forager navi-
gation between the food and nest. When trails are laid
constitutively there is a requirement for a more sophisti-
cated mechanism of redirecting foragers, one that does
not rely on trails decaying until they are undetectable.
When trails are important for orientation the modulation
of trail pheromone concentration enables foragers to se-
lect the most rewarding of two or more food sources. As
we have shown, this modulation can be achieved by two
different means, depending on whether the trail is essen-
tial for orientation. In Pharaoh’s ant modulation of trail
strength is by a change in the intensity of trail marking
by fed workers returning to the nest. Orientation trails
will be maintained but where they lead to food the trail
pheromone concentration will be increased because of
more intense marking by fed ants. Understanding the pa-
rameters constraining pheromone trail function provides
a better appreciation of the ecological limits of this forag-
ing strategy. In determining key parameters we can also
better understand the many subtle variations on the
general mechanism of self-organized information sharing
using pheromone trails.
JACKSON & CHA
ˆLINE: ANT PHEROMONE TRAIL MARKING 469

Acknowledgments
D.J. wishes to thank BT Exact and the EPSRC for funding
this research through the project ‘Simulation of Social
Insect Communities’. We also thank Francis Ratnieks for
providing facilities and Mike Holcombe for contributing
useful ideas.
References
Aron, S., Beckers, R., Deneubourg, J. L. & Pasteels, J. M. 1993.
Memory and mass communication in the orientation of two
mass-recruiting ant species. Insectes Sociaux,40, 369e380.
Beckers, R., Deneubourg, J. L., Goss, S. & Pasteels, J. M. 1990.
Collective decision-making through food recruitment. Insects
Sociaux,27, 258e267.
Beckers, R., Deneubourg, J. L. & Goss, S. 1992. Trails and U-turns
in the selection of a path by the ant Lasius niger.Journal of Theoret-
ical Biology,159, 397e415.
Beckers, R., Deneubourg, J. L. & Goss, S. 1993. Modulation of trail-
laying in the ant Lasius niger and its role in the collective selection
of a food source. Journal of Insect Behaviour,6, 751e759.
Billen, J. P. 1987. New structural aspects of the Dufour’s and venom
glands in social insects. Naturwissenschaften,74, 340e341.
Camazine, S., Deneubourg, J., Franks, N. R., Sneyd, J., Theraulaz,
G. & Bonabeau, E. 2001. Self-organisation in Biological Systems.
Princeton, New Jersey: Princeton University Press.
Detrain, C., Natan, C. & Deneubourg, J. L. 2001. The influence of
the physical environment on the self-organised foraging patterns
of ants. Naturwissenschaften,88, 171e174.
Fourcassie
´, V. & Deneubourg, J. L. 1994. The dynamics of collec-
tive exploration and trail-formation in Monomorium pharaonis:
experiments and model. Physiological Entomology,19, 291e300.
Goss,S.,Aron,J.L.,Deneubourg,J.L.&Pasteels,J.M.1989. Self-organ-
ised shortcuts in the Argentine ant. Naturwissenschaften,76, 579e581.
Hangartner, W. 1969. Structure and variability of the individual
odor trail in Solenopsis geminata.Zeitscrift fuer Vergleichende Phys-
iologie,62, 111e120.
Hart, A. & Jackson, D. E. 2006. U-turns on ant pheromone trails.
Current Biology,16,42e43.
Ho
¨lldobler, B. & Wilson, E. O. 1990. The Ants. Cambridge, Massa-
chusetts: Belknap Press of Harvard University.
Jackson, D. E., Holcombe, M. & Ratnieks, F. L. W. 2004. Trail
geometry gives polarity to ant foraging networks. Nature,432,
907e909.
Jackson, D. E., Martin, S. J., Holcombe, M. & Ratnieks, F. L. W.
2006. Longevity and detection of persistent foraging trails in Phar-
aoh’s ants, Monomorium pharaonis (L.). Animal Behaviour,71,
351e359.
Jeanson, R., Ratnieks, F. L. W. & Deneubourg, J. L. 2003. Pher-
omone trail decay rates on different substrates in the Pharaoh’s
ant, Monomorium pharaonis.Physiological Entomology,28,
192e198.
Mailleux, A. C., Deneubourg, J. M. & Detrain, C. 2000. How do
ants assess food volume? Animal Behaviour,59, 1061e1069.
Mailleux, A. C., Deneubourg, J. M. & Detrain, C. 2003. Regula-
tion of ants’ foraging to resource productivity. Proceedings of the
Royal Society of London, Series B,270, 1609e1616.
Mailleux, A. C., Detrain, C. & Deneubourg, J. M. 2005. Triggering
and persistence of trail-laying in foragers of the ant Lasius niger.
Journal of Insect Physiology,51, 297e304.
Ritter, F. J., Bruggemann, I. E. M., Persoons, C. J., Talman, E.,
van Oosten, A. M. & Verwiel, P. E. J. 1977a. Evaluation of
social insect pheromones in pest control, with special reference
to subterranean termites and Pharaoh’s ants. In: Crop Protection
Agents eTheir Biological Evaluation (Ed. by N. R. McFarlane),
pp. 201e222. London: Academic Press.
Ritter, F. J., Bruggemann, I. E. M., Verwiel, P. E. J., Persoons,
C. J. & Talman, E. 1977b. Trail pheromone of the Pharaoh’s
ant, Monomorium pharaonis: isolation and identification of faranal,
a terpenoid related to juvenile hormone II. Tetrahedron Letters,30,
2617e2618.
Sudd, J. H. 1960. The foraging method of the Pharaoh’s ant. Animal
Behaviour,8,67e75.
Sumpter, D. J. T. & Beekman, M. 2003. From nonlinearity to opti-
mality: pheromone trail foraging by ants. Animal Behaviour,66,
273e280.
Witte, V. & Maschwitz, U. 2002. Coordination of raiding and
emigration in the ponerine army ant Leptogenys distin-
guenda: a signal analysis. Journal of Insect Behaviour,15,
195e217.
ANIMAL BEHAVIOUR, 74,3
470