1. Biological Conservation Vol. 83, No. 2, pp. 221 230, 1998
© 1998 Elsevier Science Ltd
All rights reserved. Printed in Great Britain
il PII: S0006-3207(97)00054-2 0006-3207/98 $19.00 + 0.00
ELSEVIER
THE IMPACT OF CHANGES IN IRRIGATION PRACTICES
ON THE DISTRIBUTION OF F O R A G I N G EGRETS
A N D HERONS (ARDEIDAE) IN THE RICE FIELDS OF
CENTRAL JAPAN
Simon J. Lane* & Masahiro Fujioka
Applied Ornithology Laboratory, National Agriculture Research Centre, Kannondai 3-1-1, Tsukuba City, Ibaraki 305, Japan
(Received 17 April 1996; accepted 11 November 1996)
Abstract INTRODUCTION
Traditionally rice Oryza sativa fields in Japan are irriga-
ted by diverting river water through shallow earth ditches Rice Oryza sativa (L.) fields can be important foraging
using sluices and weirs. However this 'old-style' method is habitats for egrets and herons, Ardeidae (Yamagishi et al.,
being replaced rapidly by a 'new-style' system in which 1980; Fasola, 1986; Hafner et al., 1986; Fasola and Ruiz,
water is pumped into paddy fields via taps and drained 1996; Fasola et al., 1996). In Japan, recent changes in the
into deep concrete-sided canals. Concern has been way rice paddies are irrigated have led to concerns that
expressed that the changes may cause paddy fields to rice fields have declined in value as foraging habitats for
decline in value as foraging habitats for egrets and herons, Ardeidae (Narusue and Uchida, 1993). Until recently,
Ardeidae, because their aquatic prey are unable to move water was typically supplied to paddy fields via shallow
easily into new-style fields. We tested this hypothesis by irrigation canals and earth ditches which perhaps had a
comparing numbers of Ardeidae and their prey in paired nearby river as their source and were then drained at a
old- and new-style areas at six sites in central Japan from location somewhere downstream (Imamura et al., 1993).
May to August 1995. In this 'old-style' system the water levels of the paddy
Few grey herons Ardea cinerea, great egrets Egretta fields and ditches are similar and water movement is
alba or little egrets E. garzetta were recorded at our controlled by weirs and sluices. This system provides an
sites and no differences were observed in their use of old- easy means for the movement of freshwater fish and
and new-style areas. In contrast cattle Bubulcus ibis and other aquatic animals between ditches and fields.
intermediate E. intermedia egrets were common but only However to facilitate more efficient rice production,
the latter preferred old-style areas. With respect to prey easier ways of controlling the water table in paddy fields
animals, we found greater numbers and biomasses of are required and this has brought about wholesale
.frogs, crayfish Procambarus clarkii, loach Misgurnus changes in irrigation methods in many districts since
anguillicaudatus, and other fish in old-style paddy fields, 1964 (Hasegawa and Tabuchi, 1995). These 'new-style'
and in old-style ditches we found more frogs, tadpoles and irrigation systems involve the installation of powerful
fish, but generally fewer crayfish. There were no differen- pumping stations and extensive networks of under-
ces in the numbers or biomass of terrestrial invertebrates ground water pipes. Water enters paddy fields via taps,
between the two types of paddy field. Reason for these and is drained into deep, concrete-sided canals. Clearly
observations are discussed. We suggest other topics such a system may restrict the entry of aquatic animals
which require investigation in order that effective wildlife into the rice fields. Narusue and Uchida (1993) sugges-
management strategies for rice fields may be formulated. ted that both abundance and species composition of fish
© 1998 Elsevier Science Ltd and other vertebrate and invertebrate prey of egrets and
herons are reduced in new-style systems which in turn
Keywords: amphibians, Ardeidae, fish, foraging, irriga- reduces the numbers of these birds foraging in rice
tion methods, Japan, rice fields. fields. If this is correct the consequences for wildlife
dependent on the rice field habitat are potentially severe
since 80% of the 2.70 million ha of paddy fields in
*Correspondence to: S. J. Lane. Departamento de Ecologla Japan are currently irrigated by the new-style method
Evolutiva, Museo Nacional de Ciencias Naturales, CSIC, Jos6 (Ministry of Agriculture, Forestries and Fisheries, pers.
Gutirrrez Abascal 2, 28006, Madrid, Spain comm.) and more areas are converted each year.
221

2. 222 S. J. Lane, M. Fujioka
Each year, from March to September, six Ardeidae
occur commonly in Honshu, Japan: black-crowned
night heron Nycticorax nycticoracx (L.), grey heron
Ardea cinerea (L.), great egret Egretta alba (L.), inter-
mediate (= plumed) egret E. intermedia (Wagler), little
egret E. garzetta (L.) and cattle egret Bubulcus ibis (L.)
(Brazil, 1991). In late spring mixed breeding colonies,
usually of 100-500 pairs (Egret Research Team WBSJ,
1981), form in woodland, and important foraging
habitats during the following summer include rivers and
rice paddies (Brazil, 1991) where these birds feed on
amphibians, fish, insects, and crustaceans (Hancock and
Kushlan, 1984). Here, we assess the impact of changes
in paddy field irrigation methods on the numbers and (a)
distribution of Ardeidae and their prey, by comparing
their occurrence in old-style and new-style areas at six
locations in central Japan.
STUDY AREA
The study was conducted in southwestern Ibaraki pre-
fecture from May to August 1995. The region, which
forms part of the Kanto Plain around Tokyo, is bor-
dered by Mt Tsukuba (36°lYN 140°061E) to the north,
Lake Kasumigaura (36°02'N 140°231E) to the east, and
to the south the River Tone runs broadly west to east at
about 35°50~N. A predominant activity in the region is
farming and large areas are used for rice production. Of
1413 km 2 in southern Ibaraki, 26-6% are rice paddies.
Rice fields are flooded to a depth of 10-15 cm in mid
to late April and the rice seedlings are planted in April
and the first week of May. Towards the end of June the
water is drained from the fields to promote subsurface
drainage, a process necessary to harden the soil in
readiness for the entry of combine harvesters (Hase-
gawa and Tabuchi, 1995). Water is usually returned to,
and drained from, the fields intermittently in the fol-
lowing weeks, although in some cases fields remain dry
until the harvest. The precise timing of these processes
varies from farm to farm. Since the average farm size is
only 0-5-2 ha and may consist of fields of 0-3 ha or less, (b)
different approaches to water management by individual
farmers creates a mosaic of flooded, partially flooded Fig. 1. (a) Shallowearth ditch and (b) deep concrete-sidedditch
and dry paddy fields over large areas from late June typical of old-style and new-stylepaddy fields respectively.
until the harvest in late August.
In January 1995 we searched the region for discrete
areas of old-style and new-style paddy fields equal to or outflow pipes which end 0-5 m above the water level of
greater than 30 ha. Our criteria for old-style paddy the ditch, and the lower half of the ditches should have
fields were that all ditches should be shallow (that is the vertical concrete sides (Fig. l(b)).
water level should be only a few cm above or below that We found a total of six suitable locations with old-
of the paddies which they service), ditch banks should style areas and adjacent or nearby new-style areas
be of earth rather than concrete, and water should pre- (Table 1). The Hojo and Makabe locations are situated
ferably enter the fields passively and not by pumping at the foot of Mt Tsukuba, Miho and Sakuragawa are
(Fig. 1(a)). In new-style paddies, however, water should on the shore of Lake Kasumigaura, and Tone and Isobe
be pumped and enter through taps, water level in dit- are close to large rivers (Fig. 2). Our original criteria
ches should be at least 1 m below that of the paddy were met except at the Makabe new-style and at Sakur-
fields, water should be drained from the fields through agawa old-style areas where water was supplied by a

3. Rice field irrigation methods and foraging egret distribution 223
Table 1. Name and description of study sites in central Japan
Location Area (ha) Water management
Paddy fields a Other areas b Total Source c and supply a Drain e
Tone
Old-style 32-2 4-0 36.2 R,P S
New-style 32.9 0 32.9 R,T D
Miho
Old-style 27.9 4.4 32.3 L,P S
New-style 29.2 0.4 29.6 L,T D
Sakuragawa
Old-style 78.1 1.9 80-0 L,PT S
New-style 66.8 1.6 68.4 L,T D
Hojo
Old-style 41.6 3.1 44.7 M,P S
New-style 42.8 3-2 46.0 ?,T D
Isobe
Old-style 48.5 3.5 52.0 L,P S
New-style 40.7 2.8 43.5 ?,T D
Makabe
Old-style 39.2 4.3 43-5 M,P S
New-style 37.8 0 37.8 ?,PT D
alncluding areas of associated drains and access roads.
blncluding fallow and abandoned paddy fields, lotus fields, and vegetable gardens.
cWater source: R, River Tone; M, run off from Mount Tsukuba; L, Lake Kasumigaura; ?, unclear.
dSupply: P, enters paddy fields passively usually via shallow earth ditches; T, pumped and enters via taps; PT, mix of taps and
passive flow system.
eDrainage: S, water is drained from paddy fields into shallow earth-banked ditches; D, water is drained into deep concrete-sided
ditches.
IBARAKI PREF.
1
PACIFIC
M T . TSUKUBA
LAKE KASUMIGAURA
N o
~,VERTONE ,~
CHIBA PREF.
Fig. 2. The locations of the six study areas in southwestern Ibaraki prefecture, Honshu, Japan. Location 1, Makabe; 2, Isobe;
3, Hojo; 4, Miho; 5, Sakuragawa; and 6, Tone.

4. 224 S. J. Lane, M. Fujioka
combination of passive flow and pumping, and at Miho The Mondori traps did not sample frogs effectively.
where both old- and new-style areas were slightly smaller Frogs in ditches were therefore quantified by sweep
than 30 ha. At all six locations restructuring of paddy sampling. Two dip-nets (area 0.08 m 2) were placed 3 m
fields to the new-style irrigation system took place at apart on the bottom and towards the edge of a ditch
least 4 years prior to this study. For each site the areas and then brought rapidly together in order to trap the
of paddy fields, fallow paddies (estimated to have not animals. This process was repeated five times at distan-
been used for 1 or 2 years), abandoned paddies (domi- ces of at least 4 m along the ditch and no nearer than
nated by reeds or rank vegetation) and vegetable gar- 5 m to the trap.
dens were established (Table 1).
Pad@fields
METHODS At each sampling point, frogs and tadpoles in paddy
fields were sampled along 10 m of the field edge to a dis-
Numbers and distribution of egrets and herons tance of 0.6 m into the field. These prey were searched for
The six study locations were each visited on eight occa- by eye, counted, and as many as possible were caught
sions at intervals of 14 to 16 days from the beginning of with a small dip-net and weighed. In many instances not
May until the end of August. On each visit S. J. L. all prey were caught and so biomass for each species was
walked at approximately 3 km h -~ along fixed routes estimated by taking the mean dry weight of those caught
(mean length = 2.7km, n = 12) in old- and new-style and multiplying this by the total numbers seen.
areas such that no part of any area was further than Loach were too difficult to locate by eye, and crayfish
220 m from the survey paths. All egrets and herons in were too quick to be captured, and so to sample these
the area were identified and recorded. The surveys were species in paddy fields baited Mondori traps (35cm ×
conducted between 09:00 h and 13:00 h and whether the 18 cm x 9cm) were placed within arm's length of the
new- or old-style area was surveyed first was determined bank and left for 5-7 h.
randomly. To avoid biases associated with poor Neither searching nor trapping adequately sampled
weather, surveys were not conducted during heavy rain. fish (other than loach) in paddy fields. Fish were there-
Since black-crowned night herons are predominantly fore quantified by night-time sampling. The six study
nocturnal our survey method was inadequate for this locations were each visited on one occasion during the
species. Although a few individuals were recorded in the period 31 May-12 June and surveys were conducted
study the results are not presented. after nightfall between 19:30h and 01:00h. Along the
field edges in new- and old-style areas, ten 50 m × 0.6 m
Numbers and distribution of prey animals belts equidistant from one another along the daytime
Prey abundance in paddy fields, ditches and on adjacent egret survey route were searched with a powerful flash-
vegetated banks in new and old-style areas were esti- light (200,000 cw). Fish observed were counted, identi-
mated on the same day as the surveys except for the first fied when possible and data were tape-recorded. After
visits at the beginning of May when no data were col- mid-June, the rice crop became too tall and dense for
lected. We used a combination of techniques, subject to this method to be effective.
the constraints that we were not permitted to enter
paddy fields nor walk upon some of the smaller earth Invertebrates on grass banks
banks. In each old- and new-style area, prey were sam- Terrestrial invertebrates, an important food source for
pled at four fixed points equidistant from one another some egrets, were sampled on the grassy banks between
along each of the survey routes. For all sampling meth- paddy fields by sweep-net sampling. At each sampling
ods in ditches and paddy fields, aquatic prey items cap- point a sweep net (radius 18 cm) was passed four times
tured of over 1 cm in length were identified (at least to through the vegetation a little above ground level while
genus), counted, towel-dried, weighed on a portable walking along the bank. The process was repeated five
electronic balance and released. Five to ten individuals times at intervals of about 4 m. Prey greater than 1 cm in
of each prey species were collected and analysed for length were removed to the laboratory, identified (at
water content so that all prey sampled could be expres- least to order), counted, dried, and weighed.
sed in terms of dry weight as well as numbers.
Statistical analyses
Ditchs Egret densities are given as the number of birds recor-
At each sampling point, Mondori traps (60cm long x ded per 50 ha. For prey items, data from the four sam-
24 cm width x 24 cm) were used to estimate abundances pling points were averaged to give a mean. The results
of American crayfish Procambarus clarkii (Girard) are expressed as: numbers and biomass (g dry weight) of
loach Misgurnus anguillicaudatus (Cantor), other fish, prey items per 15 m of ditch and per 10 m of paddy field
and tadpoles. The traps were baited with a 1 : 1 mixture edge for sweeping in ditches and searching in paddy
of flour and a commercially available insect powder fields; numbers and biomass per 5 h for the trapping
and set in ditches for periods of 5-7h before being methods; and numbers and biomass per 20 sweeps for
removed and examined. terrestrial invertebrates. Differences between old- and

5. Rice fieM irrigation methods and foraging egret distribution 225
new-style paddy fields are assessed using repeated mea- Numbers and distribution of prey animals
sures ANOVAs after log~ (x + 1) transformations.
Ditches
RESULTS Differences were detected, between new- and old-style
ditches, in the numbers caught of some egret prey ani-
Numbers and distribution of egrets and herons mals with the exception of loach (Fig. 4).
The numbers of great egret and little egret observed at In the case of crayfish more individuals were always
our six study locations were small and no differences in caught in new-style ditches at five of the six locations,
the way these species used old- and new-style paddy the exception being at Tone. When the crayfish data for
fields were apparent (Fig. 3). Following Zar (1984), and Tone were omitted from the analysis the result became
assuming a significance level of 0.05 and an 80% prob- s i g n i f i c a n t (F1.2s = 8.21, p = 0.008) which suggests there
ability of detecting a difference, the minimum detectable are more crayfish in new-style ditches.
differences between the population means in these ana- The numbers of frogs, tadpoles, and fish (other than
lyses were 0.6 individuals 50ha -I in the case of great loach) were all significantly greater in old-style ditches.
egrets and 1.7 individuals 50 ha -1 for little egrets. In the case of the frogs the data are combined for the
In contrast, cattle egret and intermediate egret were three species captured (Japanese tree frog Hylajaponica
more abundant (Fig. 3). In the case of the cattle egret, (Gfinther), Japanese brown frog Rana japonica (Grin-
no differences were observed in the distribution of the ther) and Tokyo daruma pond frog R. porosa (Cope)),
species between old- and new-style areas. However, the whilst in the case of fish the data are combined for at
intermediate egret showed a preference for old-style least eight species: field gudgeon Gnathopogon elongatus
areas at the beginning of May, when no individuals (Sauvage); topmouth gudgeon Pseudorasbora parva
were sighted in any of the new-style areas, and again (Temminck et Schlegel); silver crucian carp Carassius
from July onwards. In these analyses the minimum langsdorfii (L.); rose bitterling Rhodeus ocellatus (Kner);
detectable differences were 6.8 and 3.4 individuals 50 ha-1 high-eyes Oryzias latipes (Temminck et Schlegel); com-
for cattle egret and intermediate egret respectively. mon freshwater goby Rhinogobius brunneus (Temminck
Only one grey heron was observed during the whole et Schlegel); snakehead Channa argus (Cantor) and
period of the study. bluegill Lepomis macrochirus (Rafinesque).
Great egret Little egret • Old-style
20 20
[] New-style
16 16
12 12
8 8
4
-"~ 0 ~ ~,..T. 0
o MayMayJun Jun Jui Jul AugAug MayMayJun Jun Jul Jul AugAug
tt~
,13
20 20 t Intermediate egret
2; /
16 1 6 ~
12 12
8 8
4 4
0 0
MayMayJun Jun Jul Jul AugAug MayMayJun Jun Jul Jul AugAug
Fig. 3. Numbers (50ha ~) of Ardeidae in old- and new-style paddy fields from the first half of May to the second half of August
1995. Bars represent means + standard errors (n = 6 old- and 6 new-style areas). With respect to type of paddy field, repeated
measures ANOVAs gave: great egret F~.4o = 0.14, p = 0-71; cattle egret Fl.40 = 0.55, p = 0.46; little egret Fj.40 = 0.26, p = 0.72; and
intermediate egret F~.40= 9.33, p = 0.004.

6. 226 S. J. Lane, M. Fujioka
~=i]Crayfish [] New-style 21t ayfi~[~ [
12 ~~~i~~'l
10
~¢~ Cr sh~ f~~~~
"Old-style ~.~ 201
~,~ 16
Z- - 0 i
~ 0
_
MayMayJun Jun .lul Jul AugAug MayMayJun Jun Jul Jul AugAug
~ 71453
• i Loach
~ ="~~,16
1.28L°ach~~]tl t
0.1 ~
"~ '~
Z
~ 0.4
0 MayMayJun Jun Jul Jul AugAug 0.0MayMayJunJun .lul Jul AugAug
. rogs
~dl. ~ 0 . 4 0 ' 8 ~
1.2 ~-o
0 . 0 ~
X'layMay.lun Jun Jtd Jul AugAug MayMayJun Jun Jul Jul AugAug
~1 Tadpoles 012 Tadpoles
:='1J.i
tt31
o
~0.4
0.8
().0
tt~ 0.08
r~ 0.06
~ 0.04
(/.02
().()0
MayMay.ltm.lun Jul .lul AugAug MayMayJun Jun .[ul Jul AugAug
-, 16Fish l
t ~= 1.2Fish
]
~ 1 . 0 ~ ~ Y
:o6
Z~ (I.4-
4 0.2
0 MayMayJun Jun Jul Jtll AtlgAtlg
f).O MayMayJun Jun Jul Jul AugAug
Fig. 4. Numbers and biomass (g dw):1: standard errors of prey animals in old- and new-style ditches between May and August 1995
(n = 6 old- and 6 new-style areas). Note that no data were collected in the first half of May (--), and the sampling methods differed
so units are not the same: for frogs (captures 15 m -1 of ditch), but for American crayfish Procambarus clarkii, loach Misgurnus
anguillicaudatus, tadpoles and fish (captures 5 h-l). With respect to ditch type, repeated measures ANOVAs gave:
Numbers Biomass
F1.35 p FI.35 P
Crayfish 1.57 0.22 1.23 0-78
Loach 0.05 0.82 1.51 0.23
Frogs 6-84 0.01 0.01 0-92
Tadpoles 4-76 0.04 1.70 0.20
Fish 5.41 0.03 4.02 0-05
None of the interaction terms with time were significant.

7. Rice field irrigation methods and foraging egret distribution 227
In terms of biomass, the data for all categories old-style paddy fields can be attributed to differences in
showed broadly the same trends as observed for ditch structure. Many species of fish are known to use
numbers although the differences between ditch type paddy fields to spawn including loach, field gudgeon
were only significant in the case of fish and in the case and rose bitterling (Saitoh et al., 1988). In the old-style
of crayfish when data from Tone were excluded system spawning fish can easily enter paddy fields from
(F1.28 = 4-57, p = 0.04). adjacent ditches because the water levels in fields and
ditches are similar. Such movement is next to impossible
Paddy fields for fish in the new-style areas where the water level in
Consistent significant differences were found in the drainage ditches may be more than 1 m below that of
numbers and biomass of most prey animals between the paddy fields. When the old-style fields are drained in
paddy fields irrigated by the old- and new-style systems midsummer both fry and adult fish can pass easily into
(Fig. 5). old-style ditches which could account for the marked
More crayfish, loach and frogs (three species com- increase in numbers of fish in these ditches in July and
bined) were found in old-style areas, but no significant August.
differences were recorded for tadpoles. Why there should be more crayfish in new-style
Night sampling in paddy fields also revealed more fish ditches is not clear, although the difference was often
(excluding loach) in old-style paddy fields (Table 2). It substantial and, with few exceptions, was consistent
was difficult to identify the fish because individuals seen throughout the study. Despite occurring in greater
were small (lengths exceeded 4 cm very rarely); however, numbers in new-style ditches crayfish were only rarely
species probably included topmouth gudgeon, field found in new-style paddy fields, suggesting that move-
gudgeon and silver crucian carp. ment between ditch and paddy field is impeded.
All three species of frog observed spawn in paddy
Invertebrates on grass banks fields (Maeda and Matsui, 1989), but more frogs
Of all terrestrial invertebrates sampled, 80.1% were were found in the old-style areas. Both R. porosa and
either slant-faced locust Aerida turrita (L.) or rice-field R. japonica may favour old-style areas because they
locust Oxya spp. The remainder were: Orthoptera require easy access from ditches to fields where they
( Metrioptera hime (Furukawa); Conocephalus sp.; spawn at night before returning to the ditches during
Ducetia japonica (Thunb.) Gryllulus spp.; Gastrimargus the day. Neither species is able to climb, so the concrete
marmoratus (D. H.); Mantoda (Mantidae); Diptera walls of ditches in new-style areas may prevent access to
(Tipula sp.); Lepidoptera (Pieris rapae (L.); Ricania paddy fields. H. japonica is arboreal and because of its
japonica (Melichar); Zizeeria maha (Koll); and Odonata climbing ability is more able to cross new ditches to
(Agrionidae, Lestidae, Libelluridae). For all orthopter- spawn in new-style paddy fields. Most tadpoles in new-
ans combined, no differences were found in the number style paddy fields were therefore probably H. japonica.
or biomass caught between old- and new-style areas Because of increasing concern about the worldwide
(Fig. 6). decline of amphibian populations (Wake, 1991; Blaustein
and Wake, 1995) a more detailed account of our find-
ings with respect to frogs is reported elsewhere (Fujioka
DISCUSSION and Lane, 1997).
We evaluated the hypotheses that old-style paddy fields Numbers and distribution of egrets and herons
support greater numbers of foraging egrets and herons Sightings of great egrets were infrequent, and only one
than new-style areas and hold more aquatic prey items, grey heron was observed in our study areas, although
such as amphibians, crustaceans and fishes (Narusue both species were seen foraging at nearby rivers and at
and Uchida, 1993). In summary, our study showed no Lake Kasumigaura. The diets of great egrets and grey
differences in the numbers of great egret, little egret or herons can include crustaceans, amphibians and insects,
cattle egret between the two paddy fields types, but but in the main their preference is for larger fish (Voisin,
more intermediate egrets were observed in the old-style 1991; Tojo, 1996). We encountered fish greater than
areas except in June. With respect to aquatic prey, most 10cm in length only rarely in this study which suggests
categories in which a significant difference occurred that neither old- or new-style rice fields in our area are
were more abundant in the old-style fields and ditches. optimal foraging habitats for these species.
The principal exception was that of crayfish which were Little egrets are often the most numerous day herons
more numerous in the new- style ditches. There were no in rice fields in Japan (Yamagishi et al., 1980; Tojo,
differences in the numbers or biomass of terrestrial 1996), but rather few were observed in this study. The
invertebrates between the two types of paddy field. data probably reflect the true status of this species in
southern Ibaraki, since a cursory inspection of selec-
Numbers and distribution of prey animals ted colonies in the area in 1992 indicated little egrets
The greater numbers of fish seen during the June night to be less abundant than cattle or intermediate egrets
surveys, and of loach trapped throughout the study, in (M. Fujioka, personal observation).

8. 228 S. J. Lane, M. Fujioka
N o differences were observed in the n u m b e r s o f little (Kosugi, 1960; Voisin, 1991) and the few numbers o f
egrets feeding in the old- and new-style areas, but the these prey in new-style p a d d y fields and ditches suggest
small n u m b e r seen makes it difficult to assess with con- the change could be important. Results f r o m elsewhere
fidence whether the changes in irrigation m e t h o d s affect in J a p a n support this contention. In Saitama Prefecture
this species. The preference o f little egrets for small fish (west o f T o k y o ) , where little egrets are numerous, m o r e
• Old-style
1.4] Crayfish
1.21Crayfish
0.8
{).6
0.4
i [] New-style
Z:
12o
tt~ 1.0
0.8
i 0.6
0.4
(1.2
0.0
MayMay Jun Jun Jtll Jul AugAug MayMay Jun Jun Jul Jul Aug Aug
,7.
4 Loach Loach
.l: 3 0.8
0.6
0.4
.,., 0 . 2
"7.
0
10
8
6
4
Maytvlay Jun Jun Jul Ju[ Aug Aug
Frogs
~
201162
Frogs
0.0
0.8
MayMay Jun Jun Jul Jul Aug Aug
2 •. 0.4
0 0.0
0.4~
MayMay Jun Jun Jul Jul AugAug NlayMayJun Jun Jul Jul AugAug
16! Tadpoles ":' 0"6 t Tadpoles
~12
~ (1.51 T
8 0.3
0.2
4-
Z 0
• 0.1
a.0
M~yMay Jun .lun Jul .lul AugAug MayMay Jun Jun Jul Jul Aug Aug
Fig. 5. Numbers and biomass (g dw) + standard errors of prey animals in paddy fields in old- and new-style areas between May
and August 1995 (n = 6 old- and 6 new-style areas). Note that no data were collected in the first half of May (--), and the sampling
methods differed so units are not the same: for American crayfish Procambarus clarkii and loach Misgurnus anguillicaudatus
(captures 5 h 1), but for frogs and tadpoles (captures 10m-~). With respect to paddy field type, repeated measures ANOVAs gave:
Numbers Biomass
Fj.35 P Fi.35 P
Crayfish 16.15 < 0.001 10.45 0.003
Loach 12.92 0.001 9.01 0.005
Frogs 17.90 < 0.001 14.16 < 0.001
Tadpoles 0.00 0.96 I. 11 0-30
None of the interaction terms with time were significant, except for numbers of tadpoles in paddy fields: F6.35 = 2.91, p = 0.02.

9. Rice field irrigation methods and foraging egret distribution 229
Table 2. Numbers of fish 50m -I ( ± standard errors, n = 6) data were not collected at this time, but it is probable
along the edges of old- and new-style paddy fields counted during that after the paddy fields are flooded by the farmers in
night-sampling in June late April prey are able to enter old-style paddy fields
Old-style New-style Statistic b more rapidly. However, from late May until the end of
June intermediate egrets occurred in similar numbers in
Fish ~ 2.88 + 2.25 0.07 ± 0.05 p < 0.05
both old- and new-style areas. At this time they foraged
aExcluding loach. almost exclusively in flooded fields, taking tadpoles
bWilcoxon matched pairs test. which were common in both types of paddy field.
In July and August intermediate egrets were once
again recorded with greater frequency in old-style areas
were observed in an area of old-style paddy fields than where frogs and loach, which are important prey for
in a nearby new-style area (Narusue and Uehida, 1993). this species (Kosugi, 1960), were more common. From
Cattle egrets were the most frequently sighted species the end of June onwards, the rapidly growing rice crop
at our study sites. They are perhaps the most terrestrial increased in both height and density and precluded
of all Ardeidae and their most important food items are intermediate egrets from entering the fields. Although
insects with orthopterans often making up the bulk of foraging in rice fields did occur in both new- and old-style
the diet (Kosugi, 1960; Voisin, 1991). We detected no areas at the field boundaries, our observations suggest
difference in the way cattle egrets used the old- and new- that the shallow, old-style ditches became an important
style areas, and the foods hunted by this species were foraging microhabitat but that the deep ditches of new-
common in both paddy field types. In May and June, style areas tended to be avoided (unpublished data).
when terrestrial invertebrates, especially grasshoppers, Compared with new-style ditches, fish were more abun-
were scarce, cattle egrets were observed foraging for dant in old-style ditches at this time and the water tended
tadpoles which were common in both old- and new-style to be less turbid, which presumably made hunting easier.
paddy fields at this time. Later, in July and August,
orthopterans became common on the grass banks of Wildlife conservation in rice fields and the need for
both old- and new-style areas where cattle egrets were future research
often seen hunting. If managed sympathetically, temporary wetlands such
Intermediate egrets were counted almost as frequently as rice fields are potentially valuable areas for nature
as cattle egrets. However, the species is listed as 'rare' in conservation. Effective management of new-style paddy
the Red Data Book of Japan (Environment Agency, fields for egrets requires a better understanding of the
1991) on the basis that it has declined in numbers at importance of rice fields as foraging areas and also of
many colonies since the 1960s (Nakamura, 1984). A the relationships between egrets and their prey.
survey of 79 heronries by the Wild Bird Society of Japan Future research should concentrate on at least five
(Egret Research Team WBSJ, 1981) found that less than basic aspects in order to formulate effective wildlife
3% of nesting birds were intermediate egrets. Why management options. (1) The relative importance to
intermediate egrets should be more common in southern Ardeidae of paddy fields as foraging habitats compared
Ibaraki than elsewhere in Japan is not known. to rivers, lakes and ponds needs to be assessed widely
Intermediate egrets showed a preference for old-style throughout Japan. (2) The most important prey animals
paddy fields, although this was not consistent through- for egrets in paddy fields need to be established. (3) The
out the season. At the beginning of May all intermedi- ecology and life histories of the important prey species
ate egrets seen were recorded in old-style areas. Prey and in particular how these are affected by changes in
• Old-style
"7
Orthopterans [] New-style .7~, 0 . 6 Orthopterans
~' 4 o.4
3 e-~ 0.3
0~
-~ 2 ~ 0.2
I ~0.1
-- -,,-i
0 m 0.0
MayMayJun Jun Jul Jul AugAug MayMayJun Jun Jul Jul AugAug
Fig. 6. Numbers and biomass (g dw) ± standard errors of orthopterans on the grass banks of old- and new-style paddy fields
between May and August 1995 (n = 6 old- and 6 new-style areas). Note that no data were collected in the first half of May (--).
With respect to paddy field type, repeated measures ANOVAs gave results of FI 35 = 0.00, p = 0-94 and of F135 = 0-10, p = 0.76 for
numbers and biomass respectively

10. 230 S . J . Lane, M. Fujioka
paddy field structure and different water management Fasola, M., Canova, L. and Saino, N. (1996) Rice fields sup-
regimes must be understood. (4) The relative impor- port a large portion of herons breeding in the Mediterra-
tance of different microhabitats for Ardeidae within the nean region. Colonial Waterbirds 19, 129-134.
Fujioka, M. and Lane, S. J. (1997) The impact of changing
paddy field environment, such as fallow paddy fields, irrigation practices in rice fields on frog populations of the
shallow and deep ditches, grass banks, flooded, wet and Kanto Plain, central Japan. Ecological Research 12, 101-
dry paddy fields and grass banks, and how each of these ll2.
can best be managed, needs to be established. (5) The Hafner, H., Dugan, P. J. and Boy, V. (1986) Use of artificial
impact of other farming practices on wildlife such as the and natural wetlands as feeding sites by little egrets (Egretta
garzetta L.) in the Camargue, Southern France. Colonial
applications of herbicides and pesticides must be assessed. Waterbirds 9, 149-154.
Hancock, J. and Kushlan, J. (1984) The Herons Handbook.
Harper and Row, New York.
ACKNOWLEDGEMENTS Hasegawa, S. and Tabuchi, T. (1995) Well facilitated
paddy fields in Japan. In Paddy fields of the world, ed.
T. Tabuchi and S. Hasegawa, pp. 103-124. Japanese Society
We thank Megumi Koga and Pat V~zquez for help with of Irrigation, Drainage and Reclamation Engineering,
field work; the many farmers who allowed us to work Tokyo.
on their lands; Kazuo Nakamura (K. N.) for identifying Imamura, N., Tsuboi, N. and Odagiri, T. (1993) Japanese
some invertebrate prey items; and Chris Elphick, farm structure: trends and projections. In Japanese and
Hiroyoshi Higuchi, K. N., Masae Narusue, Navjot American agriculture, ed. L. Tweeten, C. L. Dishon, W. S.
Chern, N. Imamura and M. Morishima, pp. 51-69. West-
Sodhi, Eiichiro Urano and an anonymous referee for view Press, Boulder, CO.
commenting on the manuscript. S. J. L. was hosted by Kosugi, A. (1960) On the food habits of some herons. Journal
K. N. at the Laboratory of Applied Ornithology and of the Yamashina Institute for Ornithology 2, 89-99 (in
held a fellowship from the Science and Technology Japanese with English summary).
Agency of Japan following nomination by the European Maeda, N. and Matsui, M. (1989) Frogs and Toads of Japan.
Bun-Ichi Sogo Shuppan, Tokyo.
Commission Directorate General for Science, Research Nakamura, K. (1984) An introduction to egrets. Yacho 49(8),
and Development. M. F. is a guest scientist of the Cen- 14-17 (in Japanese).
tre for Ecological Research, Kyoto University. Narusue, M. and Uchida, H. (1993) The effect of structural
changes of paddy fields on foraging egrets. Strix 12, 121-
130 (in Japanese with English summary).
REFERENCES Saitoh, K., Katano, O. and Koizumi, A. (1988) Movement
and spawning of several freshwater fishes in temporary
Blaustein, A. R. and Wake, D. B. (1995) The puzzle of waters around paddy fields. Japanese Journal of Ecology 38,
declining amphibian populations. Scientific American 242, 35--47 (in Japanese with English summary).
56--61. Tojo, H. (1996) Habitat selection, foraging behaviour and
Brazil, M. (1991). The birds of Japan. Christopher Helm, prey of five heron species in Japan. Japanese Journal of
London. Ornithology 45, 141-158.
Egret Research Team, WBSJ (1981). [Survey of egret colonies Voisin, C. (1991) The Herons of Europe. T and D Poyser,
and roosts in Japan by questionnaire.] Wild Bird Society of London.
Japan, Tokyo (in Japanese). Wake, D. B. (1991) Declining amphibian populations. Science,
Environment Agency (1991). [Red data book (vertebrates) of N.Y. 253, 860.
Japan.] Environment Agency of Japan, Tokyo (in Japanese). Yamagishi, S., Inoue, Y. and Komeda, S. (1980) Distribution
Fasola, M. (1986) Resource use of foraging herons in agricul- of colonies and roosts and feeding dispersal of the little a
tural and nonagricultural habitats in Italy. Colonial Water- cattle egrets (Egretta garzetta and Bubulcus ibis) in Nara
birds 9, 139-148. Basin, Japan. Tori 29, 69-85 (in Japanese with English
Fasola, M. and Ruiz, X. (1996) The value of rice fields as summary).
substitutes for natural wetlands for waterbirds in the Medi- Zar, J. H. (1984) Biostatistical Analysis. Prentice Hall, New
terranean region. Colonial Waterbirds 19, 122-128. Jersey.