Our reference: AGEE 4997 P-authorquery-v9
AUTHOR QUERY FORM
Journal: AGEE Please e-mail your responses and any corrections to:
Article Number: 4997 E-mail: corrections.esch@elsevier.thomsondigital.com
Dear Author,
Please check your proof carefully and mark all corrections at the appropriate place in the proof (e.g., by using on-screen
annotation in the PDF file) or compile them in a separate list. Note: if you opt to annotate the file with software other than
Adobe Reader then please also highlight the appropriate place in the PDF file. To ensure fast publication of your paper please
return your corrections within 48 hours.
For correction or revision of any artwork, please consult http://www.elsevier.com/artworkinstructions.
Any queries or remarks that have arisen during the processing of your manuscript are listed below and highlighted by flags in
the proof. Click on the ‘Q’ link to go to the location in the proof.
Location in Query / Remark: click on the Q link to go
article Please insert your reply or correction at the corresponding line in the proof
Reference(s) given here were noted in the reference list but are missing from the text – please position
each reference in the text or delete it from the list.
Q1 Please confirm that given names and surnames have been identified correctly.
Q2 The affiliation ‘a’ has been split into two different affiliations, as two different country names were
provided for the same affiliation. Please check, and correct if necessary.
Q3 Please check the address for the corresponding author that has been added here, and correct if necessary.
Q4 Please check all the keywords and amend if necessary.
Q5 Reference "Brouder et al., 2014" has been changed to "Brouder and Gomez-Macpherson, 2014" in the
text to match the name in the reference list, please check and amend if necessary.
Q6 “Your article is registered as a regular item and is being processed for inclusion in a regular issue of
the journal. If this is NOT correct and your article belongs to a Special Issue/Collection please contact
j.shanmugam@elsevier.com immediately prior to returning your corrections.”
Q7 Reference "Wood, 1988" has been changed to "Wood and Cowie, 1988" in the text to match the name
in the reference list, please check and amend if necessary.
Q8 Please check the edit made in these sentence and amend if necessary.
Q9 Please check the edit made in these sentence and amend if necessary.
Q10 Reference "Wood, 1996" is cited in the text but not provided in the reference list, Please position these
reference in the list or, alternatively, delete it.
Q11 Uncited references: This section comprises references that occur in the reference list but not in the body
of the text. Please position each reference in the text or, alternatively, delete it. Any reference not dealt
with will be retained in this section.
Q12 One or more sponsor names and the sponsor country identifier may have been edited to a standard format
that enables better searching and identification of your article. Please check and correct if necessary.
Q13 Please supply the volume no. for Ref. "Garnett et al., 2013".
Q14 Please provide an update for reference “Kihara et al.., in press”.
Q15 Please check the correctness of Ref. "Maniania et al., 2001" and amend if necessary.
Page 1 of 2
Location in Query / Remark: click on the Q link to go
article Please insert your reply or correction at the corresponding line in the proof
Q16 Please supply the name of the publisher for Ref. "Swift and Bignell, 2001".
Q17 Please check the presentation of all the tables and amend if necessary.
Q18 Please provide the citation of P value (*, **) in Table 3 as the corresponding table note text have been
mentioned.
Q19 Please provide the citation of P value (*, ** and ***) in Table 4 as the corresponding table note text
have been mentioned.
Q20 Please provide the citation of P value (*, **) in Table 5 as the corresponding table note text have been
mentioned.
Please check this box or indicate your approval if
you have no corrections to make to the PDF file
Thank you for your assistance.
1 Highlights
Agriculture, Ecosystems and Environment xxx (2015) pp. xxx–xxx
Exclusion of soil macrofauna did not affect soil quality but
increased crop yields in a sub-humid tropical maize-based
system
B.K. Paul c,*, B. Vanlauwed, M. Hoogmoed 1, T.T. Hurisso c,2, T. Ndabamenye c,3, Y. Terano c, J. Six, F.O. Ayuke 4, M.M.
Pulleman c

 Termites were dominant soil macrofauna irrespective of tillage and residue management.

 Termite diversity was low and grass/residue feeding species were dominant.

 Soil macrofauna did not affect soil C content nor soil aggregate stability.

 Maize and soybean yields were higher when soil macrofauna were excluded.
Agriculture, Ecosystems and Environment xxx (2015) xxx–xxx
G Model
AGEE 4997 1
Contents lists available at ScienceDirect
Agriculture, Ecosystems and Environment
journal homepage: www.elsev ier .com/locate /agee
1 Exclusion of soil macrofauna did not affect soil quality but increased
2 crop yields in a sub-humid tropical maize-based system
3 B.K. Paul a,b,c,*Q1 , B. Vanlauwe d, M. Hoogmoed c,1, T.T. Hurisso c,2, T. Ndabamenye c,3,
4 Y. Terano c, J. Six e, F.O. Ayuke c,4, M.M. Pulleman a,b,c
5 aCIAT (International Center forQ2 Tropical Agriculture), Cali, Colombia
6 bCIAT (International Center for Tropical Agriculture), P.O. Box 823-00621, Nairobi, Kenya
7 cWageningen University, Department of Soil Quality, P.O. Box 47, 6700AA Wageningen, the Netherlands
d IITA (International Institute of Tropical Agriculture), Natural Resource Management Research Area, P.O. Box 30772-00100, Nairobi, Kenya
e ETH-Zurich, Swiss Federal Institute of Technology, Department of Environmental Systems Science, Zurich 8092, Switzerland
A R T I C L E I N F O
Article history:
Received 26 December 2014
Received in revised form 27 March 2015
Accepted 2 April 2015
Available online xxx
Keywords:
Termites
Earthworms
Ants
Tillage
Residue retention
Soil structure
Conservation agriculture
Western Kenya
Sub-Saharan Africa
A B S T R A C T
Soil macrofauna such as earthworms and termites are involved in key ecosystem functions and thus
considered important for sustainable intensification of crop production. However, their contribution to
tropical soil and crop performance, as well as relations with agricultural management (e.g., Conservation
Agriculture), are not well understood. This study aimed to quantify soil macrofauna and its impact on soil
aggregation, soil carbon and crop yields in a maize–soybean system under tropical sub-humid conditions.
A field trial was established in Western Kenya in 2003 with tillage and residue retention as independent
factors. A macrofauna exclusion experiment was superimposed in 2005 through regular insecticide
applications, and measurements were taken from 2005 to 2012. Termites were the most abundant
macrofauna group comprising 61% of total macrofauna numbers followed by ants (20%), while few
earthworms were present (5%). Insecticide application significantly reduced termites (by 86 and 62%)
and earthworms (by 100 and 88%) at 0–15 and 15–30 cm soil depth respectively. Termite diversity was
low, with all species belonging to the family of Macrotermitinae which feed on wood, leaf litter and dead/
dry grass. Seven years of macrofauna exclusion did not affect soil aggregation or carbon contents, which
might be explained by the low residue retention and the nesting and feeding behavior of the dominant
termites present. Macrofauna exclusion resulted in 34% higher maize grain yield and 22% higher soybean
grain yield, indicating that pest damage – probably including termites – overruled any potentially
beneficial impact of soil macrofauna. Results contrast with previous studies on the effects of termites on
plant growth, which were mostly conducted in (semi-) arid regions. Future research should contribute to
sustainable management strategies that reduce detrimental impact due to dominance of potential pest
species while conserving soil macrofaunaQ4 diversity and their beneficial functions in agroecosystems.
ã 2015 Published by Elsevier B.V.
81. Introduction
9Feeding a rapidly growing human population while preserving
10environmental sustainability results in unprecedented challenges
11for agriculture and natural resources. Sustainable intensification is
12especially urgent in Sub-Saharan Africa where soil degradation and
13food insecurity are most pressing, and smallholder farmers are
14challenged by scarcity of resources including organic and inorganic
15fertilizers (Godfray et al., 2010; Garnett et al., 2013). It is of crucial
16importance that management of agricultural soils enhances and
17sustains soil fertility and resource use efficiency, based on a better
18understanding of ecosystem services (Beare et al., 1997; Brussaard
19et al., 2010). Management practices involving minimum soil
20disturbance, organic soil cover and crop diversification –
* CorrespondingQ3 author at: Wageningen University, Department of Soil Quality, P.
O. Box 47, 6700AA Wageningen, the Netherlands. Tel.: +254 208632800;
fax: +254 208632001.
E-mail address: B.Paul@cgiar.org (B.K. Paul).
1 Monash University, School of Biological Sciences, Clayton, VIC 3800, Australia.
2 University of Wyoming, College of Agriculture and Natural Resources, Laramie,
WY 82071-3354, USA.
3 Rwanda Agriculture Board (RAB), Infrastructure and Mechanization Directorate,
P.O. Box 5016 Kigali, Rwanda.
4 University of Nairobi, Department of Land Resource Management and
Agricultural Technology, P.O. Box 30197, 00100 Nairobi, Kenya.
http://dx.doi.org/10.1016/j.agee.2015.04.001
0167-8809/ã 2015 Published by Elsevier B.V.
Agriculture, Ecosystems and Environment xxx (2015) xxx–xxx
G Model
AGEE 4997 1–11
Please cite this article in press as: Paul, B.K., et al., Exclusion of soil macrofauna did not affect soil quality but increased crop yields in a sub-
humid tropical maize-based system. Agric. Ecosyst. Environ. (2015), http://dx.doi.org/10.1016/j.agee.2015.04.001
Contents lists available at ScienceDirect
Agriculture, Ecosystems and Environment
journal homepage: www.elsev ier .com/locate /agee
21 collectively known as Conservation Agriculture (CA) – are widely
22 promoted in Sub-Saharan Africa. CA has been shown to stimulate
23 soil macrofauna, which, in turn, can lead to enhanced soil
24 aggregation and therefore C storage, reduced runoff and erosion,
25 improved nutrient and water use efficiency, and ultimately stable
26 crop yields. These potentialQ5 impacts however vary with specific
27 agro-ecologies (Erenstein, 2003; Hobbs, 2007; Palm et al., 2014;
28 Brouder and Gomez-Macpherson, 2014; Corbeels et al., 2014).Q6
29 A wide range of different soil macrofauna provide ecosystem
30 services including soil organic matter turnover, nutrient cycling,
31 soil structure formation and pest and disease control (Lavelle et al.,
32 1997; Beare et al., 1997; Brussaard, 2012). Of key interest are soil
33 ecosystem engineers such as earthworms, termites and, to lesser
34 extent ants. Through bioturbation they incorporate plant litter and
35 crop residues into the soil, thereby modifying biological, chemical
36 and physical soil processes that affect the flow of energy and
37 material, and modify the habitat of other soil biota (Jones et al.,
38 1994; Lavelle et al., 1997; Pulleman et al., 2012). The impact of soil
39 ecosystem engineers on soil quality is partly mediated through
40 their effects on soil organic matter incorporation and soil
41 aggregation (Six et al., 2004). Stable soil aggregates can physically
42 protect soil organic matter against rapid decomposition (Six et al.,
43 2000) and reduce soil erosion (Barthes and Roose, 2002). It has
44 been shown that the biogenic structures produced by earthworms
45 and termites, such as casts and sheetings, can differ from bulk soil
46 in organo-physical composition and be enriched in carbon and
47 nutrients, suggesting protection of organic matter against rapid
48 mineralization (Fall et al., 2001; Mora et al., 2003; Pulleman et al.,
49 2005; Bossuyt et al., 2005). The capacity of earthworms to
50 incorporate organic matter into the soil and improve soil
51 aggregation has been widely investigated (Lee, 1985; Lavelle
52 et al., 1997; Six et al., 2004), although effects on C mineralization
53 versus C stabilization are still a matter of debate (Lubbers et al.,
54 2013). It has also been shown that earthworm abundance is
55 generally higher in no-tillage systems due to the lack of mechanical
56 soil disturbance (Chan, 2001; Shuster and Edwards, 2003;
57 Castellanos-Navarrete et al., 2012).
58 Termites are considered the dominant soil ecosystem engineers
59 in tropical (semi)-arid areas, whereas earthworms occur widely in
60 (semi-) humid climates, both tropical and temperate (Lal, 1988;
61 Lobry de Bruyn and Conacher, 1990). Termites are well known for
62 their role in organic matter breakdown and modification of soil
63 properties. They construct a variety of organo-mineral structures
64 that result from intestinal transit (casts) or are mixed and
65 impregnated with saliva and are used to construct mounds, nests,
66 galleries and surface sheetings (Lobry de Bruyn and Conacher,
67 1990; Fall et al., 2001; Mora et al., 2003). Termites can mold up to
68 1300 kg ha
1 of soil annually (Kooyman and Onck, 1987) and it has
69 been suggested that their biogenic structures constitute microsites
70 that protect organic C against rapid mineralization (Mora et al.,
71 2003). Termites are classified on the basis of their food choice,
72 feeding habits and nesting behavior, ranging from soil feeders that
73 occur in the soil profile and feed on organic matter associated with
74 mineral soil, wood feeders that feed on wood and excavate
75 galleries in larger items of wood litter, and litter feeders that forage
76 for leaf litter, dry standing grass stems and small woody items
77 (Swift and Bignell, 2001; Eggleton and Bignell, 1995; Wood, 1996).
78 Similarly, ants modify the soil through foraging and nest building
79 although their impact on soil properties is generally less important
80 compared with earthworms and termites (Jungerius et al., 1999;
81 Lobry de Bruyn and Conacher, 1990).
82 A number of studies focusing on natural Savanna ecosystems in
83 Australia and West Africa have reported beneficial effect of
84 termites on soil porosity, water infiltration, nutrient uptake and
85 plant cover or biomass, demonstrating their capacity to rehabil-
86 itate degraded and crusted soils (Sarr et al., 2001; Dawes, 2010;
87Mando and Brussaard, 1999; Ouedraogo et al., 2004). In Kenya, the
88role of termites and ants in the formation of the microgranular
89structure of Ferralsols was studied by Jungerius et al. (1999),
90whereas Fall et al. (2001) compared the effects of different termite
91feeding groups on soil organic matter and aggregate fractions in
92West African semi-arid Savanna. Few studies exist, however, that
93have evaluated the effects of termites or ants on agricultural soils
94(Kooyman and Onck, 1987; Lobry de Bruyn and Conacher, 1990;
95Evans et al., 2011), as research on termites in Q7agricultural systems
96has historically focused on their pest role (Wood and Cowie, 1988;
97Black and Okwakol, 1997; Bignell, 2006). Positive effects of tropical
98soil macrofauna on crop yields have been demonstrated experi-
99mentally in a limited number of studies, again in (semi-) arid
100climates in West Africa (Ouedraogo et al., 2006, 2007) and West
101Australia (Evans et al., 2011), where low rainfall and poor surface
102structure strongly constrain crop production. The impact of soil
103macrofauna on soil structural properties, soil C and crop
104performance in (sub) humid climates, as well as their response
105to soil tillage and crop residue management in CA systems is
106largely unclear (Giller et al., 2009).
107The overall aim of our study was to quantify the contribution of
108soil macrofauna (earthworms, termites, and ants) to soil aggrega-
109tion, soil C and crop productivity as a function of different tillage
110and residue management under sub-humid climatic conditions.
111The hypotheses tested were:
(i) 112Soil tillage and residue removal negatively affect the abun-
113dance of soil macrofauna;
(ii) 114Soil macrofauna increase stable soil aggregation and soil C;
(iii) 115Soil macrofauna increase crop yields through positive effects
116on soil quality.
1172. Materials and methods
1182.1. Site description
119This study was conducted in an existing long-term conservation
120tillage trial in Nyabeda, Siaya district, Nyanza province in Western
121Kenya. The site is located at an altitude of 1420 m asl, latitude 0 060
122N and longitude 34 240E, and the slope is less than 2%. A mean
123annual rainfall of 1800 mm (sub-humid) is distributed over two
124rainy seasons: the long rainy season lasts from March until August
125and the short rainy season from September until January (Fig. 1).
126The soil was identified as a Ferralsol (FAO, 1998), with five
127distinctive soil horizons. The upper soil horizon (0–8 cm) had 58%
128clay, 24% sand, and 18% silt. Soil chemical characteristics of the
129same soil layer included pH (H2O) 5.1, total N 0.16%, total C 2%, Bray
130P 8 ppm, Olsen P 8.1 mg kg
1. The second soil horizon (9–40 cm)
131had 72% clay, 14% sand, 14% silt, and pH (H2O) 5.6, total N 0.15%,
132total C 1.6%, Bray P 1 ppm, Olsen P 2.7 mg kg
1 (Nic Jelinski, 2014,
133unpublished data).
1342.2. Experimental design and trial management
135The field experiment was established in March 2003, and has
136been managed by researchers and technicians of the International
137Center for Tropical Agriculture (CIAT). The trial was set up in a
138randomized block design (n = 4) with tillage and crop residue
139retention as main factors. Each factor had two levels: conventional
140tillage (+T) or reduced tillage (-T) and residue retention (+R) or
141residue removal (-R). Individual plots measured 7 4.5 m. The crop
142rotation since trial establishment has consisted of soybean (Glycine
143max L.) during short rains and maize (Zea mays L.) during long
144rains. All plots were fertilized with 60 kg ha
1 N (urea), 60 kg ha
1 P
145(Triple Super Phosphate) and 60 kg ha
1 K (Muriate of Potash) per
2 B.K. Paul et al. / Agriculture, Ecosystems and Environment xxx (2015) xxx–xxx
G Model
AGEE 4997 1–11
Please cite this article in press as: Paul, B.K., et al., Exclusion of soil macrofauna did not affect soil quality but increased crop yields in a sub-
humid tropical maize-based system. Agric. Ecosyst. Environ. (2015), http://dx.doi.org/10.1016/j.agee.2015.04.001
146 growing season. To control stem borer, 5 kg ha
1 of granulated
147 Bulldock (beta-cyfluthrin) was applied in the funnel of the maize
148 plants during the 5th week after planting in all treatments. Under
149 conventional tillage (+T), the seedbed was prepared by hand
150 hoeing to 15 cm soil depth. Weeding was performed three times
151 per season, using the hand hoe. Under reduced tillage (-T), a 3 cm
152 deep seedbed was prepared with the hand hoe. Weeding was
153 performed three times per season by hand pulling until the long
154 rainy season of 2009. Thereafter, herbicides (glyphosate and 2,4-
155 dichlorophenoxyacetic acid) have been applied to all reduced
156 tillage treatments before planting and subsequent weeding was
157 done by hand pulling. Maize residues were collected after harvest,
158 dried, chopped and stored during the dry season. At the time of
159 soybean planting residues were reapplied at a rate of 2 t ha
1 (+R).
160 Since soybeans drop leaves prior to grain maturity, soybean
161 residues (leaves and stems) always remained in the field after
162 harvest, irrespective of treatment. These soybean residues were
163 then either incorporated (+T) or remained at the soil surface (-T).
164 Further details regarding maize and soybean planting and fertilizer
165 application, are described in Paul et al. (2013).
166 A macrofauna exclusion experiment was superimposed on the
167 tillage and residue management trial as a split-plot factor starting
168 in the short rainy season of 2005. Subplots (2  4.5 m) with
169 macrofauna exclusion (+Exc) were created through the application
170 of insecticides, just before planting and every three weeks until
171 harvest. Dursban was used at 0.8 l ha
1, with 400 g ha
1 of active
172 ingredient (chlorophyrifos) to eliminate termite activity. Endoco-
173 ton was applied at 0.9l ha
1,Q8 with 450 g ha
1 of active ingredient
174 (endosulfan) to exclude earthworm activity. These rates are based
175 on effect levels determined by Ouédraogo et al. (2004). By lowering
176the spraying tip to approximately 10 cm above the soil surface,
177contact with crops was minimized. The subplots with and without
178macrofauna exclusion were separated by 30 cm high metal sheets
179that were installed 15 cm into the soil to avoid contamination with
180insecticides in the 
Exc treatment.
1812.3. Soil macrofauna abundance and taxonomic diversity
182Soil macrofauna was sampled in the short rainy season of 2005
183(12 weeks after planting), long rainy season of 2006 (15 weeks after
184planting), and the short rainy season of 2009 (6 weeks after
185planting), corresponding Q9to the 6th, 7th and 12th cropping season
186after the tillage trial establishment. We used monolith sampling as
187described by Bignell et al. (2008) and Anderson and Ingram (1993).
188One soil monolith measuring 25 cm  25 cm  30 cm depth was
189randomly taken in each replicate plot (n = 4). The extracted soil was
190divided into two depth increments (0–15 and 15–30 cm) and
191macrofauna was collected in the field by hand sorting on plastic
192trays. Macrofauna were killed in 70% ethanol, and then stored in
193sealed vials, whereas earthworms were first killed in 70% ethanol,
194then fixed in 4% formaldehyde before being transported to the
195laboratory for identification and enumeration. Macrofauna abun-
196dance was determined in all three years and classified according to
197the following main groups: earthworms (Oligochaeta), termites
198(Isoptera), ants (Formicidae) and other macrofauna. Complete
199identification to the genus level was done for the 2006 samples
200only. Specimens were identified in the Department of Invertebrate
201Zoology of the Nairobi National Museum, using keys and reference
202specimens in their collections.
2032.4. Soil and crop analyses
204During the long rainy seasons of 2006 (15 weeks after planting)
205and 2008 (14 weeks after planting) and in 2012 (4 weeks before
206maize planting), undisturbed soil samples were taken from all
207treatments (n = 4) at 0–15 cm and 15–30 cm soil depth for soil
208aggregate analysis. Soil samples in field moist conditions were pre-
209sieved over an 8 mm sieve and air dried before wet sieving into four
210aggregate size fractions as described by Elliot (1986): (a) large
211macroaggregates (LM; >2000 mm), (b) small macroaggregates (SM;
212250–2000 mm), (c) microaggregates (Mi; 53–250 mm), (d) silt and
213clay sized particles (SC; 53 mm). 80 g of air-dried soil was evenly
214spread on a 2 mm sieve, which was placed in a recipient filled with
215deionized water and left to slake. After 5 min, the sieve was
216manually moved up and down 50 times in 2 min. The procedure
217was repeated passing the material on to a 250 mm and 53 mm sieve.
218Soil aggregates retrieved at each sieve were carefully backwashed
219into beakers, oven-dried at 60 C for 48 h, weighed back and stored
220for C and N analysis. SC was calculated from the total volume of the
221suspension and the volume of the subsample. Mean weight
222diameter (MWD) was determined as the sum of the weighted
223mean diameters of all fraction classes (Van Bavel, 1950). In the
224same years total soil C and N was performed: composite samples
225consisting of 4 cores per subplot were taken at 0–15 cm and
22615–30 cm soil depth. All samples were oven-dried at 60 C for 48 h,
227ground and weighed, and then sent to UC Davis, California, USA.
228Total C and N were determined with a Dumas combustion method,
229using a PDZ Europa ANCA-GSL elemental analyzer (Sercon Ltd.,
230Cheshire, UK). In 2012, a more detailed sampling took place. C and
231N sampling was done as previously but depths were 0–5 cm,
2325–15 cm and 15–30 cm. Bulk density was measured for the same
233depths: two undisturbed samples (5 cm diameter metal rings;
234100 cm3 volume) per subplot were taken at 0–5 cm, 7.5–12.5 cm
235and 20–25 cm. Samples were dried at 105 C for 24 h prior to
236weighing. Soil C contents for 0–15 cm soil depth in 2012 were
237calculated from the 0–15 to 5–15 cm soil depth data taking into
Fig. 1. Seasonal cumulative rainfall (mm) from 2006 to 2012 during the long rainy
season (a) and short rainy season (b) in Nyabeda, Western Kenya. Maize was grown
in the long rainy season (March/April–August) and soybean during the short rainy
season (August/September–January/February). Cumulative rainfall is adjusted to
planting and harvest dates per year.
B.K. Paul et al. / Agriculture, Ecosystems and Environment xxx (2015) xxx–xxx 3
G Model
AGEE 4997 1–11
Please cite this article in press as: Paul, B.K., et al., Exclusion of soil macrofauna did not affect soil quality but increased crop yields in a sub-
humid tropical maize-based system. Agric. Ecosyst. Environ. (2015), http://dx.doi.org/10.1016/j.agee.2015.04.001
238 account bulk density (weighted averages). Maize and soybean
239 biomass and grain yields were determined as described in Paul
240 et al. (2013) and reported on an oven-dry basis. Soybean yield data
241 for the 2005 short rainy season were omitted from our data set
242 because no distinction was made between the +/
 insecticide
243 treatments during harvest. Daily rainfall was measured with a
244 rainfall gauge in the experimental field.
245 2.5. Statistical analyses
246 Analysis of variance (ANOVA) was carried out with R Studio
247 Version 0.97.449 (R Core Team, 2013). A Linear Mixed Model was
248 fitted by REML using the lmerTest package. ANOVA is calculated
249 based on Satterthwaite’s approximation (Kuznetsova et al., 2014).
250 Tillage, residue management, and insecticide application and their
251 interactions were included in the model as fixed factors, and effects
252 were tested on macrofauna abundance, soil aggregate stability, soil
253 C and crop yields. Block and year were defined as random factors,
254 and the autocorrelations of plot (tillage and residue treatments)
255 and subplot (insecticide treatment) were accounted for. Macro-
256 fauna and soil aggregate data were analysed independently for two
257 soil depths (0–15 cm and 15–30 cm) and soil C for three depths (0–
258 5 cm, 0–15 cm and 15–30 cm). Macrofauna abundance data were
259 square-root transformed before analysis to fit ANOVA assumption
260of normal data distribution. One monolith that contained a
261subterranean nest was removed from the dataset (year 2009; +T
262+ R treatment; block 4; 0–15 cm depth 156 termites; 15–30 cm
263depth 2602 termites) and replaced by a missing value. Means are
264presented with standard errors. A P-value  0.05 was considered
265significant.
2663. Results
2673.1. Macrofauna abundance and species as affected by management
268The most numerous macrofauna group present across all years
269and depths were termites, with an average abundance of 204 ind.
270m
2 across all years, tillage and residue treatments. This
271constitutes 61% of all macrofauna. Ants were on average 68 ind.
272m
2 or 20% of all macrofauna, earthworms were present in very
273low numbers (15 ind. m
2; 5%) and other macrofauna constituted
27446 ind. m
2 or 14% of total macrofauna (Table 1). Tillage and
275residue treatments did not affect macrofauna abundance except a
276significant residue-insecticide interaction effect for other macro-
277fauna at 0–15 cm (P = 0.049, Table 1). All genera of termites found
278(mostly Pseudacanthotermes and Microtermes) belong to the
279subfamily of Macrotermitinae, which are fungus growers and
280feed on wood, leaf litter and dead/dry grass (Table 2). Earthworms
Table 1
Soil macrofauna abundances Q17(earthworms, termites, ants, others) in no m2 across three sampling times (2005, 2006 and 2009) at 0–15 cm and 15–30 cm soil depths.
Soil macrofauna abundance (no m
2)
Tillage Residue Exclusion Termites Earthworms Ants Other fauna
0–15 cm soil depth

 
 
 77.3 (22.2) 1.3 (1.3) 18.0 (8.4) 8.0 (3.1)

 
 + 13.1 (7.1) 0 (0) 49.3 (45.1) 6.7 (4.1)

 + – 36.0 (11.1) 4.0 (4.0) 82.7 (54.5) 28.0 (10.8)

 + + 2.7 (2.7) 0 (0) 0 (0) 8.0 (6.7)
+ – 
 106.7 (45.9) 6.7 (3.7) 32.0 (29.1) 12.0 (4.9)
+ 
 + 9.3 (4.1) 0 (0) 21.3 (18.5) 14.7 (6.9)
+ + 
 75.3 (26.0) 5.3 (4.1) 28.0 (23.8) 18.7 (4.3)
+ + + 18.7 (14.8) 0 (0) 12.0 (12.0) 12.0 (7.4)
S.V.
F-ratio p-value F-ratio p-value F-ratio p-value F-ratio p-value
Tillage 0.63 0.431 1.18 0.293 0.48 0.491 0.66 0.417
Residue 1.43 0.235 0.01 0.919 0.12 0.724 1.85 0.178
Exclusion 31.05 <0.001*** 6.99 0.017* 2.78 0.099 6.58 0.012*
Til  Res 1.26 0.264 0.27 0.611 0.01 0.910 0.18 0.672
Til  Exc 0.04 0.850 1.19 0.291 0.79 0.376 0.37 0.542
Res  Exc 0.17 0.684 0.01 0.919 1.91 0.170 3.98 0.049*
Til  Res  Exc 0.17 0.685 0.27 0.609 0.86 0.357 0.01 0.916
15–30 cm soil depth

 
 
 170.7 (56.8) 5.3 (5.3) 17.3 (8.2) 16.0 (8.8)

 
 + 104.0 (55.4) 0 (0) 2.0 (1.4) 13.3 (6.5)

 + 
 50.7 (18.7) 8.0 (6.7) 12.0 (8.1) 32.0 (18.0)

 + + 53.3 (37.4) 1.3 (1.3) 1.3 (1.3) 13.3 (8.1)
+ 
 
 196.7 (58.9) 21.3 (9.1) 30.7 (16.6) 34.0 (19.5)
+ 
 + 26.7 (10.5) 1.3 (1.3) 1.3 (1.3) 14.7 (6.9)
+ + 
 104.0 (55.4) 8.0 (4.2) 51.3 (33.4) 35.3 (15.1)
+ + + 13.3 (6.2) 2.7 (1.8) 16 (14.6) 20.0 (12.4)
S.V.
F-ratio p-value F-ratio p-value F-ratio p-value F-ratio p-value
Tillage 2.23 0.161 2.68 0.127 1.78 0.195 1.00 0.319
Residue 2.58 0.134 0.03 0.855 0.08 0.775 0.46 0.498
Exclusion 4.91 0.046* 10.23 0.002** 8.63 0.007** 3.76 0.055
Til  Res 2.11 0.172 0.76 0.400 1.15 0.294 0.07 0.792
Til  Exc 0.49 0.497 1.99 0.163 0.57 0.457 0.16 0.690
Res  Exc 0.08 0.774 1.21 0.274 0.08 0.775 0.45 0.504
Til  Res  Exc 0.40 0.538 1.44 0.234 0.03 0.857 0.51 0.478
Treatments refer to combinations of reduced tillage (
) and conventional tillage (+); residue removal (
) and residue retention (+); without macrofauna exclusion (
) and
with macrofauna exclusion (+). Mean values are indicated with standard errors in parenthesis. S.V. means source of variation. Levels of significance indicate single and
interactive effects of tillage, residue and macrofauna exclusion over all three years. P values refer to the following levels of significance: *<0.05, **<0.01, ***<0.001.
4 B.K. Paul et al. / Agriculture, Ecosystems and Environment xxx (2015) xxx–xxx
G Model
AGEE 4997 1–11
Please cite this article in press as: Paul, B.K., et al., Exclusion of soil macrofauna did not affect soil quality but increased crop yields in a sub-
humid tropical maize-based system. Agric. Ecosyst. Environ. (2015), http://dx.doi.org/10.1016/j.agee.2015.04.001
281 included individuals of the epigeic Pareudrilinae (family Eudrili-
282 dae), and juveniles which could not be identified. Ants were
283 represented by 10 different genera including several well-known
284 predators such as Dorylus and Hypoponera that have been reported
285 to attack termites. Other macrofauna included mainly Isopoda and
286 Araneae, as well as Coleoptera and Lepidoptera larvae, Hemiptera,
287 Chilopoda, Diplopoda and Dermaptera (Table 2).
288 3.2. Macrofauna exclusion efficacy
289 Insecticide application reduced termite numbers by 86% (0–
290 15 cm soil depth; P < 0.001) and 62% (15–30 cm, P = 0.046, Table 1).
291 Exclusion efficacy was higher for earthworms with 100% at
292 0–15 cm (P = 0.017) and 88% at 15–30 cm soil depth (P = 0.002).
293 Insecticide was also effective for ants and other macrofauna,
294 excluding 49% (P = 0.099) and 38% (P = 0.012) at 0–15 cm and 81%
295 (P = 0.007) and 48% (P = 0.055) at 15–30 cm, respectively (Table 1).
296 3.3. Soil aggregate fractions, bulk density and soil carbon
297 Macrofauna exclusion did not have a significant effect on soil
298 aggregate fractions and mean weight diameter (MWD) at 0–15 and
299 15–30 cm soil depth. The only marginal effect was seen in a smaller
300 amount of SM fraction at 15–30 cm soil depth when macrofauna
301 was excluded (P = 0.058, Table 3), but this was not reflected in any
302 of the other size fractions. However, soil aggregate stability was
303 strongly affected by soil tillage at 0–15 cm, decreasing LM by 49%
304 (P < 0.001), and increasing Mi and SC by 29% (P < 0.001) and 45%
305 (P < 0.001) respectively. This resulted in an overall 29% lower MWD
306 under conventional tillage compared to reduced tillage (P < 0.001,
307 Table 3). Likewise, soil macrofauna exclusion did not affect bulk
308 density as measured in 2012. Bulk density ranged from 1.02/
309 1.05 g cm3 (
ins/+ins) at 0–5 cm to 1.12/1.07 g cm3 at 5–15 cm to
310 1.07/1.06 g cm3 at 15–30 cm (data not presented). Macrofauna
311 exclusion did not significantly affect soil C content at any soil
312 depth. Highest soil C content was measured under reduced tillage
313 with residue retention (-T + R 21.25 mg g
1) and lowest soil C under
314reduced tillage without residue retention (-T-R 18.33 mg g
1)
315(Table 4).
3163.4. Crop yields
317Macrofauna exclusion resulted in 34% higher maize grain yields
318across all years (P < 0.001, Fig. 2, Table 5). A marginally significant
319interactive effect of tillage and residue on maize yields was found
320(P = 0.067; Table 5). No tillage without residue retention (-T-R)
321resulted in the lowest maize yields (4.61 t ha
1) when compared
322with the other tillage and residue treatments which ranged
323between 5.12 and 5.54 t ha
1 (Fig. 2, Table 5).
324Macrofauna exclusion increased soybean grain yields by 22%
325(P < 0.001, Fig. 3, Table 5). Tillage and residue showed a marginally
326significant interaction effect (P = 0.051) resulting in lowest soybean
327grain yields under no tillage without residue retention (-T-R;
328922 kg ha
1) when compared to the other treatments which
329ranged between 1053 kg ha
1 (+T + R) and 1111 kg ha
1 (+T-R, Fig. 3,
330Table 5).
3314. Discussion
3324.1. Termite dominance and diversity, and insecticide exclusion
333efficiency
334Termites were the dominant group of soil macrofauna, followed
335by ants. Earthworms were present in very low numbers with an
336average density of 15 ind. m
2, despite the sub-humid climate.
337These earthworm numbers are in line with Ayuke et al. (2011) who
338sampled different treatments of the same field experiment during
339the 2007 long rainy season and reported earthworm densities
340ranging from 0 to 11 ind. m
2 in arable plots compared to 117 ind.
341m
2 in an adjacent long-term fallow. Such results suggest that
342higher densities of earthworms can be found in sub-humid tropical
343climates, but that their densities are strongly affected by land use
344and soil management. The same study also showed that although
345adjacent long-term fallow had higher termite densities (475 m
2;
346Ayuke et al., 2011) than the arable treatments in this study
347(204 m
2), their overall abundance was still relatively high in the
348arable plots.
349Insecticides proved to be effective in eliminating the main
350target organisms – termites and earthworms – and to a lesser
351extent reduced the abundances of ants and other macrofauna.
352Results underline that insecticide applications indeed resulted in
353the desired macrofauna exclusion, showing a considerable
354reduction in both termites (by 86 and 62%) and earthworms (by
355100 and 88%) at 0–15 and 15–30 cm depth. Exclusion efficiencies
356were higher in the upper soil layer than at 15–30 cm depth since
357insecticides were applied at the soil surface and both Dursban and
358Endocoton readily adhere to soil particles. Soil macrofauna
359exclusion through insecticide application was proven to be
360effective in 2005, 2006 and 2009 and it was assumed to be
361effective until the end of the study period in 2012.
362Termite diversity at the study site was low, only including
363genera of the subfamily of Macrotermitinae, namely Microtermes
364and Pseudacanthotermes. Although the identification was done
365only for samples collected in 2006, these finding corroborate with
366data from the same trial based on a combination of monolith and
367transect samplings (Ayuke et al., 2011), which showed the same
368genera with a strong dominance of Pseudacanthotermes. Pseuda-
369canthotermes feed on litter, grass, and even live maize plants by
370covering them with their sheetings after which litter is carried to
371their nests. Microtermes feed both on litter and woody materials by
372hollowing out mostly dead, but sometimes live plant stems and
373twigs, and entering Q10plant roots (Wood, 1996). Microtermes and
374Pseudacanthotermes are both fungus growing and have deep and
Table 2
Soil macrofauna identification for 2006 sampling (0–15 and 15–30 cm soil depths).
Macrofauna Class or Order Family Subfamily Genus
Termites Isoptera Termitidae Macrotermitinae Microtermes
Pseudacanthotermes
Others
Earthworms Oligochaeta Eudrilidae Pareudrilinae
Juveniles
(not
identified)
Ants Hymenoptera Formicidae Dolichoderinae Tecnomyrmex
Dorylinae Dorylus
Formicinae Lepsiota
Myrmicinae Acanthomyrmex
Carebara
Rhoptromyrmex
Ponerinae Cryptopone
Euponera
Harpegnathos
Hypoponera
Others Araneae
Chilopoda
Coleoptera
larvae
Dermaptera Forficulidae
Polyzonida
(Diplopoda)
Hemiptera
Isopoda
Lepidoptera
larvae
B.K. Paul et al. / Agriculture, Ecosystems and Environment xxx (2015) xxx–xxx 5
G Model
AGEE 4997 1–11
Please cite this article in press as: Paul, B.K., et al., Exclusion of soil macrofauna did not affect soil quality but increased crop yields in a sub-
humid tropical maize-based system. Agric. Ecosyst. Environ. (2015), http://dx.doi.org/10.1016/j.agee.2015.04.001
Table 4
Soil C (mg g
1) at 0–5 cm (2012 only), 0–15 cm Q19and 15–30 cm (average for 2006, 2008 and 2012) soil depths.
Soil C (mg g
1)
Tillage Residue Exclusion 0–5 cm 0–15 cm 15–30 cm

 
 
 18.47 (0.67) 19.91 (0.30) 16.33 (0.39)

 
 + 18.18 (0.43) 19.57 (0.40) 16.76 (0.44)

 + 
 21.20 (0.97) 21.02 (0.53) 18.25 (0.42)

 + + 22.30 (1.35) 20.74 (0.44) 18.69 (0.53)
+ 
 
 19.47 (0.47) 20.12 (0.28) 18.31 (0.45)
+ 
 + 19.48 (0.50) 20.36 (0.44) 18.71 (0.44)
+ + 
 19.54 (0.57) 20.51 (0.34) 18.78 (0.37)
+ + + 19.70 (0.20) 20.44 (0.36) 18.40 (0.53)
S.V.
F-ratio p-value F-ratio p-value F-ratio p-value
Tillage 0.25 0.625 0.03 0.872 4.00 0.069
Residue 3.62 0.081 2.42 0.146 3.70 0.078
Exclusion 0.92 0.342 0.25 0.618 0.69 0.423
Til  Res 3.07 0.105 1.11 0.312 3.10 0.104
Til  Exc 9.44 0.509 1.04 0.310 0.63 0.443
Res  Exc 1.90 0.175 0.36 0.550 0.51 0.488
Til  Res  Exc 1.21 0.278 0.09 0.759 0.53 0.478
Treatments refer to combinations of reduced tillage (
) and conventional tillage (+); residue removal (
) and residue retention (+); without macrofauna exclusion (
) and
with macrofauna exclusion (+). S.V. means source of variation. Mean values are indicated with standard errors in parenthesis. Levels of significance indicate single and
interactive effects of tillage, residue and macrofauna exclusion over all three years. P values refer to the following levels of significance: *<0.05, **<0.01, ***<0.001.
Table 3
Soil aggregate fractions (g 100 g
1 soil) Q18and mean weight diameter (mm) across 2006, 2008 and 2012 at 0–15 cm and 15–30 cm soil depths.
Aggregate fractions (g 100 g
1 soil) and mean weight diameter (mm)
Tillage Residue Exclusion LM SM Mi SC MWD
0–15 cm soil depth

 
 
 15.2 (3.0) 49.9 (3.6) 24.9 (2.8) 8.0 (1.0) 1.32 (0.13)

 
 + 15.0 (2.6) 49.8 (2.9) 25.4 (2.6) 6.7 (0.6) 1.33 (0.12)

 + 
 11.8 (1.9) 52.7 (2.6) 26.0 (2.8) 7.4 (0.7) 1.21 (0.09)

 + + 11.5 (0.9) 51.9 (2.3) 26.4 (2.5) 7.7 (1.1) 1.17 (0.05)
+ 
 
 5.9 (0.9) 51.1 (2.3) 32.5 (3.4) 10.3 (1.1) 0.89 (0.06)
+ 
 + 6.6 (1.1) 46.0 (2.2) 34.3 (3.4) 10.9 (1.0) 0.87 (0.06)
+ + 
 6.5 (0.6) 48.9 (3.0) 32.9 (3.0) 11.7 (1.7) 0.89 (0.05)
+ + + 8.1 (0.9) 47.7 (2.4) 33.1 (3.0) 10.4 (1.7) 0.95 (0.05)
S.V.
F-ratio p-value F-ratio p-value F-ratio p-value F-ratio p-value F-ratio p-value
Tillage 20.41 <0.001*** 2.32 0.162 35.89 <0.001*** 13.10 <0.001*** 24.26 <0.001***
Residue 0.67 0.429 0.39 0.546 0.06 0.812 0.1 0.761 0.43 0.525
Exclusion 0.28 0.600 1.85 0.178 0.53 0.482 0.37 0.543 0.00 0.960
Til  Res 2.37 0.149 0.57 0.468 0.33 0.576 0.03 0.864 1.50 0.244
Til  Exc 0.59 0.444 1.02 0.316 0.06 0.809 0.02 0.897 0.12 0.724
Res  Exc 0.04 0.839 0.35 0.554 0.22 0.648 0.02 0.878 0.63 0.801
Til  Res  Exc 0.07 0.789 0.74 0.391 0.14 0.716 1.67 0.200 0.57 0.451
15–30 cm soil depth

 
 
 23.2 (4.4) 56.0 (3.6) 14.8 (1.3) 4.6 (0.6) 1.79 (0.18)

 
 + 26.3 (4.3) 49.5 (3.1) 16.3 (1.8) 5.9 (1.9) 1.87 (0.20)

 + 
 24.0 (3.2) 52.9 (3.3) 16.7 (1.3) 6.4 (1.7) 1.81 (0.13)

 + + 20.9 (3.6) 55.2 (3.1) 17.6 (1.8) 5.1(0.6) 1.67 (0.16)
+ 
 
 24.0 (3.0) 54.1 (3.8) 16.5 (1.5) 5.1 (0.8) 1.82 (0.12)
+ 
 + 21.9 (2.1) 50.3 (1.8) 19.6 (1.5) 5.4 (0.6) 1.67 (0.10)
+ + 
 16.9 (1.5) 56.9 (2.4) 18.6 (1.6) 7.3 (1.1) 1.49 (0.05)
+ + + 18.9 (2.4) 53.7 (1.8) 19.3 (1.9) 6.0 (0.9) 1.56 (0.12)
F-ratio p-value F-ratio p-value F-ratio p-value F-ratio p-value F-ratio p-value
Tillage 2.35 0.151 0.02 0.888 4.41 0.057 0.41 0.521 2.74 0.123
Residue 3.15 0.101 1.39 0.268 1.54 0.238 1.63 0.205 3.06 0.106
Exclusion 0.00 0.997 3.72 0.058 2.44 0.122 0.12 0.731 0.19 0.664
Til  Res 0.44 0.520 0.23 0.642 0.11 0.740 0.33 0.569 0.48 0.503
Til  Exc 0.00 0.972 0.23 0.631 0.12 0.724 0.09 0.769 0.00 0.957
Res  Exc 0.11 0.745 2.62 0.110 0.56 0.456 1.97 0.163 0.00 0.957
Til  Res  Exc 2.54 0.115 1.98 0.163 0.17 0.684 0.14 0.707 2.01 0.160
Aggregate fractions include large macroaggregates (LM; >2000 mm), small macroaggregates (SM; 250–2000 mm), micro aggregates (Mi; 53–250 mm) and silt and clay (SC;
<53 mm). Mean weight diameter (MWD) is the sum of the weighted mean diameters of all fraction classes. Treatments refer to combinations of reduced tillage (
) and
conventional tillage (+); residue removal (
) and residue retention (+); without macrofauna exclusion (
) and with macrofauna exlcusion (+). S.V. means source of variation.
Mean values are indicated with standard errors in parenthesis. Levels of significance indicate single and interactive effects of tillage, residue and macrofauna exclusion over all
three years. P values refer to the following levels of significance: *<0.05, **<0.01, ***<0.001.
6 B.K. Paul et al. / Agriculture, Ecosystems and Environment xxx (2015) xxx–xxx
G Model
AGEE 4997 1–11
Please cite this article in press as: Paul, B.K., et al., Exclusion of soil macrofauna did not affect soil quality but increased crop yields in a sub-
humid tropical maize-based system. Agric. Ecosyst. Environ. (2015), http://dx.doi.org/10.1016/j.agee.2015.04.001
375 diffuse subterranean networks. Microtermes is strictly subterra-
376 nean. A nest consists of a large number of chambers located
377 between 10 cm and 2 m below the surface, although more than 80%
378 occur between 10 and 50 cm depth (Kooyman and Onck, 1987).
379 Pseudacanthotermes also constructs conical mounds outside of the
380 cultivated field (Kooyman and Onck, 1987; Eggleton and Bignell,
381 1995).
382 4.2. Tillage and residue influence on soil macrofauna
383 Crop residues can provide an important food source for
384 decomposer soil fauna and can moderate extremes in soil moisture
385 and temperature, especially when maintained at the soil surface in
386 the absence of soil tillage. Residue retention in the form of
387 mulching has previously been shown to attract termites in humid
388 and arid parts of West Africa (Tian et al., 1993; Ouedraogo et al.,
389 2004). Similarly, past research has shown the impact of tillage on
390 the abundance of soil ecosystem engineers. Holt et al. (1993) found
391 that no-till soil maintained significantly more termite gallery
392 structures than conventionally tilled soil, whereas Castellanos-
393 Navarrete et al. (2012) showed the beneficial effects of reduced
394 tillage and surface residue retention on earthworm numbers. In
395 the present study, however, we did not detect any significant
396 effects of tillage and/or residue management on soil macrofauna
397 abundance. In case of earthworms, the overall low abundances and
398 lack of residue effects may be attributed to the relatively low
399 amounts and quality of organic matter inputs in the +R treatments
400 (2 t ha
1 yr
1 of maize stover), whereas termites may also be
401 limited by low amounts of available residues especially later on in
402 the growing season (Kihara et al., 2015in press). Macrofauna data,
403especially in case of termites, were characterized by high
404variability between replicates in space and time. Taking into
405account the spatial and temporal variations in macrofauna
406abundance and nesting and/or foraging patterns, sampling
407methods are challenged to accurately measure macrofauna
408abundances in small experimental agronomic plots (Eggleton
409and Bignell, 1995). Nevertheless, our results were corroborated by
410Kihara et al. 2015 (in press) who did not find significant differences
411in termite activity in maize with or without residue application or
412tillage, based on counting of termite sheetings 14 and 16 weeks
413after residue application. The lack of tillage effect on termite
414abundance might be attributed to the termite species found. Only
415wood and litter feeding termites were present at the study site,
416whereas soil feeding termites are the most affected by tillage
417(Black and Okwakol, 1997). Moreover, we rarely encountered
418termite nests within our monolith samples, indicating that
419termites nested below ploughing depth or outside the plots where
420they are unaffected by tillage. Pseudacanthotermes spp. were
421observed to build mounds and extensive subterranean foraging
422tunnels outside of the agricultural plots, probably a strategy to
423escape from regular tillage disturbance within the plots (Kooyman
424and Onck, 1987).
4254.3. Effect of termites on soil aggregate stability and carbon
426Having shown that earthworm abundances are very low and
427that termites are the dominant soil macrofauna at our study site,
428we can relate effects of macrofauna exclusion primarily to termite
429activities. Termites have previously been found to contribute to
430higher nutrient contents (Mando and van Rheenen, 1988; Evans
Fig. 2. Maize grain yields in t ha
1 with and without macrofauna exclusion and for tillage and residue treatments –T-R (a), 
T + R (b), +T-R (c), +T + R (d) from 2006 to 2012.
Treatments refer to combinations of reduced tillage (
T) and conventional tillage (+T); residue removal (-R) and residue retention (+R); without macrofauna exclusion (
exc)
and with macrofauna exclusion (+exc).
B.K. Paul et al. / Agriculture, Ecosystems and Environment xxx (2015) xxx–xxx 7
G Model
AGEE 4997 1–11
Please cite this article in press as: Paul, B.K., et al., Exclusion of soil macrofauna did not affect soil quality but increased crop yields in a sub-
humid tropical maize-based system. Agric. Ecosyst. Environ. (2015), http://dx.doi.org/10.1016/j.agee.2015.04.001
431 et al., 2011) and improved physical soil quality (Mando and van
432 Rheenen, 1988; Dawes, 2010). A small number of studies have
433 characterized different termite-molded soil structures collected in
434 the field and reported higher stability and/or C contents compared
435 to bulk soil (Mora et al., 2003; Kooyman and Onck, 1987; Kihara
436 et al., 2015in press), but implications of termite activities for bulk
437soil aggregate stability and soil C are not known (Kihara et al.,
4382015in press). Contrary to our hypothesis, we did not find any
439significant effect of macrofauna exclusion on aggregate stability
440nor soil C content over a period of 7 years. The absence of a positive
441effect of macrofauna on soil C might be explained by residue
442translocation by termites to mounds outside of the arable plots or
Table 5
Maize grain yields in t ha
1 (a) Q20and soybean grain yields in kg ha
1 (b) from 2006 to 2012.
Crop yields (maize t ha
1; soybean kg ha
1)
Tillage Residue Exclusion Maize Soybean

 
 
 3.87 (0.31) 729 (107)

 
 + 5.35 (0.41) 1115 (144)

 + 
 4.38 (0.30) 1019 (98)

 + + 5.87 (0.32) 1112 (126)
+ 
 
 4.80 (0.29) 1047 (86)
+ 
 + 6.27 (0.42) 1174 (119)
+ + 
 4.40 (0.25) 940 (76)
+ + + 5.88 (0.35) 1165 (98)
S.V.
F-ratio p-value F-ratio p-value
Tillage 4.38 0.058 2.10 0.149
Residue 0.06 0.803 0.39 0.531
Exclusion 72.78 <0.001*** 14.05 <0.001***
Til  Res 4.06 0.067 3.83 0.051
Til  Exc 0.00 0.998 0.33 0.564
Res  Exc 0.00 0.972 0.77 0.380
Til  Res  Exc 0.00 0.998 3.12 0.079
Treatments refer to combinations of reduced tillage (
) and conventional tillage (+); residue removal (
) and residue retention (+); without insecticide (
) and with
insecticide application (+). S.V. means source of variation. Mean values are indicated with standard errors in parenthesis. Levels of significance indicate single and interactive
effects of tillage, residue and insecticide application over all three years. P values refer to the following levels of significance: *<0.05, **<0.01, ***<0.001.
Fig. 3. Soybean grain yields in kg ha
1 with and without macrofauna exclusion and for tillage and residue treatments –T-R (a), -T + R (b), +T-R (c), +T + R (d) from 2006 to 2012.
Treatments refer to combinations of reduced tillage (-T) and conventional tillage (+T); residue removal (-R) and residue retention (+R); without macrofauna exclusion (
exc)
and with macrofauna exclusion (+exc).
8 B.K. Paul et al. / Agriculture, Ecosystems and Environment xxx (2015) xxx–xxx
G Model
AGEE 4997 1–11
Please cite this article in press as: Paul, B.K., et al., Exclusion of soil macrofauna did not affect soil quality but increased crop yields in a sub-
humid tropical maize-based system. Agric. Ecosyst. Environ. (2015), http://dx.doi.org/10.1016/j.agee.2015.04.001
443 into deep subterranean networks. Kihara et al. 2015(in press)
444 showed that in the same field trial, almost 40% of surface residue
445 disappeared within the first 4 weeks after planting, and up to 85%
446 had disappeared in 3.5 months, compared to 20% in the absence of
447 macro- and mesofauna. Similarly, studies in Burkina Faso showed
448 that soil fauna (including termites) strongly increase the rate of
449 decomposition of organic residues in semi-arid areas (Mando and
450 Brussaard, 1999; Ouedraogo et al., 2004).
451 Regarding the effects of termite activities on bulk soil aggregate
452 stability it has been shown that the characteristics of different
453 termite biogenic structures can be highly dependent on the feeding
454 group present and the origin of the soil material. Fall et al. (2001)
455 showed that mound structures from two species representing two
456 different feeding groups (soil feeders versus litter feeders) in the
457 semi-arid savanna of Senegal gave highly contrasting results. The
458 fungus-growing litter feeder Macrotermes bellicosus, who build the
459 nest from subsoil particles mixed with saliva (Wood and Cowie,
460 1988), had lower C content than the reference soil and no impact
461 on soil C and soil aggregation was recorded (Fall et al., 2001).
462 Macrotermitinae use feces exclusively for construction of fungus
463 combs (Kooyman and Onck, 1987). By contrast, the soil feeding
464 termites built their nests from feces and had a high impact on soil
465 microaggregate structure, representing 60% of the total soil mass
466 and 50% of the total carbon (Fall et al., 2001). For our research site,
467 Kihara et al. 2015(in press) showed that termite sheetings collected
468 from the soil surface of the arable plots were enriched in
469 particulate organic matter and carbon compared to bulk soil,
470 but did not show elevated aggregate stability. Results strongly
471 suggest that the activities of litter feeding termites do not increase
472 bulk soil aggregate stability in arable plots and that accumulation
473 of particulate organic matter in termite molded soil is not reflected
474 in bulk soil C contents. The low residue retention rate of 2 t ha
1 in
475 comparison to the large background soil C pool is likely to be
476 insufficient to cause a significant impact on soil C, especially when
477 large proportion of the material is exported by termites. This
478 interpretation is supported by the observation that no significant
479 effect on soil C at any soil depth was found due to residue
480 treatment, irrespective of soil tillage. However, we cannot exclude
481 with certainty that a potentially negative effect of macrofauna
482 exclusion on soil aggregation was masked by increased biomass
483 production (see Section 4.4) which would be expected to stabilize
484 soil structure. Such an effect could have only been isolated through
485 a no-plant control treatment. The negative effect of tillage on
486 aggregate stability, as previously shown in Paul et al. (2013) for the
487 period 2005–2008, was also confirmed by the additional data in
488 this paper, showing a 29% decrease in MWD due to tillage at 0–
489 15 cm soil depth.
490 4.4. Effect of termites on crop yields
491 We found a strong effect of macrofauna exclusion on crop
492 yields, resulting in 34% higher maize yield and 22% higher
493 soybean yield. The insecticides Dursban and Endocoton and
494 other insecticides have been successfully used in previous
495 studies to establish soil macrofauna exclusion plots for studying
496 the effects of soil macrofauna on soil properties and crop
497 production. However, contrary to our results, all these studies
498 found higher crop yields or plant cover with soil macrofauna,
499 and attributed this effect to soil rehabilitation and increased
500 soil porosity and water infiltration through termite activity
501 (Mando and Brussaard, 1999; Ouédraogo et al., 2004; Evans
502 et al., 2011; Dawes, 2010).
503 Based on the composition of the insecticides used in terms of
504 nutrient contents or pH we can exclude a direct effect on nutrient
505 availability to plants. Therefore possible explanations might be a
506 reduction in crop pest damage (including termites) in exclusion
507plots, and/or an indirect effect in the form of enhanced residue
508retention in the absence of termites. The first mechanisms is
509supported by the fact that identified termite species all belong to
510the family of Macrotermitinae, which is responsible for 90% of
511damage in agriculture, forestry, urban settings (Mitchell, 2002).
512We cannot exclude that other pest species than termites may have
513played a role as both Endocoton and Dursban are known to impact
514on other pest organisms. However circumstantial evidence
515suggests that termite pests are at least partially responsible.
516Besides stemborer, which was controlled for in all treatments
517through the application of Bulldock (beta-cyfluthrin), termites are
518one major maize pests in Kenya (Mainaina et al., 2001). Micro-
519termes and Pseudacanthotermes have been identified by local
520farmers as major pest species in the area (Ayuke et al., 2010), and
521have been recorded to attack maize in Ethiopia, Nigeria and Zambia
522(Rouland-Lefevre, 2011). In the southern Guinea savannah zone of
523Nigeria, virtually all maize crop damage was caused by Microtermes
524by entering roots commencing 10–12 weeks after planting and
525leading to plant lodging (Wood et al., 1980). Low species richness
526can lead to an increase in relative abundance of pest species, as the
527large majority of termite species are not pests under any
528circumstances and non-pest species of termites may compete
529with pest species for similar resources (Black and Okwakol, 1997).
530Damage in maize was higher than in soybean, indicating that non-
531indigenous crops like maize are more susceptible, presumably
532because they lack co-evolution. Indirect negative effects of
533termites on soil and moisture conservation through removing
534crop residue may also have played a role. Lowest maize and
535soybean yields were found in –T-R treatments, and rapid residue
536removal by termites can therefore quickly convert Conservation
537Agriculture (-T + R) into such unfavorable states.
5385. Conclusions
539Termites were the dominant soil macrofauna at our study site in
540sub-humid Western Kenya, while earthworm densities were
541extremely low. We did not find a significant effect of tillage nor
542crop residue management on the abundance of soil macrofauna. In
543addition, no effects of soil macrofauna on soil C content were
544observed upon macrofauna exclusion over a period of 7 years.
545Results are attributed to low residue retention rates and the
546specific feeding and nesting behavior of the termites found, which
547remove crop residues and transport them to deep subterranean
548networks or mounds outside of arable plots. Negative effects of
549tillage on aggregate stability as found previously for the same site
550were confirmed, but no relation with the presence of soil
551macrofauna was found. Further, increased crop yields in treat-
552ments that excluded soil macrofauna through insecticide applica-
553tion indicate significant crop loss due to pest problems, especially
554in maize. The low termite diversity, including termites which are
555well-known potential crop pests, supports this explanation.
556Indirect negative effects of termites on residue cover, soil and
557moisture conservation may have also played a role. Further
558research is needed to elucidate these mechanisms. Our study
559contradicts earlier work showing positive effects of termites on
560physical soil properties and crop production in (semi-) arid
561climates, suggesting a decisive influence of variations in agroeco-
562logical conditions and production limiting factors such as climate,
563soil conditions and crop type, in combination with the behavior of
564the dominant termite species present.
565Uncited references Q11
566Kihara et al. (2012), Maniania et al. (2001), Ouédraogo et al.
567(2006) and Ouédraogo et al. (2007).
B.K. Paul et al. / Agriculture, Ecosystems and Environment xxx (2015) xxx–xxx 9
G Model
AGEE 4997 1–11
Please cite this article in press as: Paul, B.K., et al., Exclusion of soil macrofauna did not affect soil quality but increased crop yields in a sub-
humid tropical maize-based system. Agric. Ecosyst. Environ. (2015), http://dx.doi.org/10.1016/j.agee.2015.04.001
568 Acknowledgements
569 We greatly appreciate the help and diligent work of all
570 responsible field and lab technicians at CIAT and UC Davis,
571 especially John Mukalama, Lukelysia Nyawira Mwangi and Wilson
572 Ngului. Thanks to Rolf Sommer, Martin Stephen Macharia, and
573 Stephen Crittenden for comments, proofreading and statistical
574 advice, and Saidou Koala who assisted in accessing yield data. This
575 study was financially supported byQ12 the Netherlands Organization
576 for ScientificResearch/Science for Global Development (NWO-
577 WOTRO) through Wageningen University (grant number
578 W01.65.219.00) and a research grant from the International
579 Atomic Energy Association (IAEA) to CIAT (grant number 14447).
580 Researchers’ time was in part funded by the CGIAR Research
581 Program on Humidtropics.
582 References
583 Anderson, J.M., Ingram, J.S., 1993. Tropical Soil Biology and Fertility: A Handbook of
584 Methods. C.A.B. International, Wallingford, UK.
585 Ayuke, F.O., Pulleman, M.M., Vanlauwe, B., de Goede, R.G.M., Six, J., Csuzdi, C.,
586 Brussaard, L., 2011. Agricultural management affects earthworm and termite
587 diversity across humid to semi-arid tropical zones. Agric. Ecosyst. Environ. 140,
588 148–154.
589 Barthes, B., Roose, E., 2002. Aggregate stability as an indicator of soil susceptibility
590 to runoff and erosion; validation at several levels. Catena 47, 133–149.
591 Beare, M.H., Reddy, M.V., Tian, G., Srivastava, S.C., 1997. Agricultural intensification,
592 soil biodiversity and agroecosystem function in the tropics: the role of
593 decomposer biota. Appl. Soil Ecol. 6, 87–108.
594 Bignell, D.E., 2006. Termites as soil engineers and soil processors. In: Koenig, H.,
595 Varma, A. (Eds.), Intestinal Microorganisms of Soil Invertebrates. Soil Biology, 6.
596 Springer, Berlin, Germany, pp. 183–219.
597 Bignell, D.E., Constantino, R., Csuzdi, C., Karyanto, A., Konate, S., et al., 2008.
598 Macrofauna. In: Moreira, F.M.S., Huising, J.E., Bignell, D.E. (Eds.), A Handbook of
599 Tropical Soil Biology: Sampling and Characterization of Below-Ground
600 Biodiversity. Earthscan, UK.
601 Black, H.I.J., Okwakol, M.J.N., 1997. Agricultural intensification, soil biodiversity and
602 agroecosystem function in the tropics: the role of termites. Appl. Soil Ecol. 6,
603 37–53.
604 Bossuyt, H., Six, J., Hendrix, P.F., 2005. Protection of soil carbon by microaggregates
605 within earthworm casts. Soil Biol. Biochem. 37, 251–258.
606 Brouder, S.M., Gomez-Macpherson, H., 2014. The impact of conservation agriculture
607 on smallholder agricultural yields: a scoping review of the evidence. Agric.
608 Ecosyst. Environ. 187, 1–10.
609 Brussaard, L., Caron, P., Campbell, B., Lipper, L., Mainka, S., Rabbinge, R., et al., 2010.
610 Reconciling biodiversity conservation and food security: scientific challenges
611 for a new agriculture. Curr. Opin. Environ. Sustainability 2, 34–42.
612 Brussaard, L., 2012. Ecosystem services provided by soil biota. In: Wall, D.H. (Ed.),
613 Oxford Handbook of Soil Ecology and Ecosystem Services. Oxford University
614 Press, Oxford.
615 Castellanos-Navarrete, A., Rodríguez-Aragonés, C., de Goede, R.G.M., Kooistra, M.J.,
616 Sayre, K.D., et al., 2012. Earthworm activity and soil structural changes under
617 conservation agriculture in central Mexico. Soil Till. Res. 123, 61–70.
618 Chan, K.Y., 2001. An overview of some tillage impacts on earthworm population
619 abundance and diversity – implications for functioning in soils. Soil Till. Res. 57,
620 179–191.
621 Corbeels, M., de Graaff, J., Ndah, T.H., Penot, E., Baudron, F., et al., 2014.
622 Understanding the impact and adoption of conservation agriculture in Africa: a
623 multi-scale analysis. Agric. Ecosyst. Environ. 187, 155–170.
624 Dawes, T.Z., 2010. Reestablishment of ecological functioning by mulching and
625 termite invasion in a degraded soil in an Australian savanna. Soil Biol. Biochem.
626 42, 1825–1834.
627 Eggleton, P., Bignell, D.E., 1995. Monitoring the response of tropical insects to
628 changes in the environment: troubles with termites. In: Harrington, R., Stork, N.
E. (Eds.), Insects in a Changing Environment. Academic Press, London, pp. 473–
629 497.
630 Elliot, E.T., 1986. Aggregate structure and carbon, nitrogen and phosphorus in native
631 and cultivated soils. Soil Sci. Soc. Am. J. 50, 627–633.
632 Erenstein, O., 2003. Smallholder conservation farming in the tropics and sub-
633 tropics: a guide to the development and dissemination of mulching with crop
634 residues and cover crops. Agric. Ecosyst. Environ. 100, 17–37.
635 Evans, T.A., Dawes, T.Z., Ward, P.R., Lo, N., 2011. Ants and termites increase crop yield
636 in a dry climate. Nat. Commun. 2, 262.
637 FAO (1998) World reference base for soil resources. ISSS-AISS-IBG, ISBN 92-5-
638 104141-5.
639 Fall, S., Brauman, A., Chotte, J.L., 2001. Comparative distribution of organic matter in
640 particle and aggregate size frations in the mounds of termites with different
641 feeding habits in Senegal: Cubitermes niokoloensis and Macrotermes bellicosus.
642 Appl. Soil Ecol. 17, 131–140.
643Garnett, T., Appleby, M.C., Balmford, A., Bateman, I.J., Benton, T.G., et al., 2013.
644Sustainable intensification in Q13agriculture: premises and policies. Science 341.
645Giller, K.E., Witter, E., Corbeels, M., Tittonell, P., 2009. Conservation agriculture and
646smallholder farming in Africa: the heretics’ view. Field Crops Res. 114, 23–34.
647Godfray, H.C.J., Beddington, J.R., Crute, I.R., Haddad, L., Lawrence, D., et al., 2010. Food
security: the challenge of feeding 9 billion people. Science 327, 812–818.
648Hobbs, P.R., 2007. Conservation agriculture: what is it and why is it important for
649future sustainable food production? J. Am. Soil Agron. 145, 127–137.
650Holt, J.A., Robertson, L.N., Radford, B.J., 1993. Effect of tillage and stubble residue
651treatments on termite activity in two Central Queensland Vertisols. Aust. J. Soil
652Res. 31, 311–317.
653Jones, C.G., Lawton, J.H., Shachak, M., 1994. Organisms as ecosystem engineers.
654Oikos 69, 373–386.
655Jungerius, P., van den Ancker, J.A., Mücher, H., 1999. The contribution of termites to
656the microgranular structure of soils on the Uasin Gishu Plateau, Kenya. Catena
65734, 349–363.
658Kihara, J., Mukalama, J., Ayuke, F.O., Njoroge, S., Waswa, B., et al., 2012. Crop and soil
659response to tillage and crop residue application in a tropical ferralsol in sub-
660humid Western Kenya. In: Bationo, A. (Ed.), Lessons Learned from Long-Term
661Soil Fertility Management Experiments in Africa. Springer, Dordrecht.
662Kihara, J., Martius, C., Bationo, A., 2015. Crop residue disappearance and macrofauna Q14
663activity in sub-humid Western Kenya. Nutr. Cycling Agroecosyst. (in press).
664Kooyman, C., Onck, R.F.M., 1987. Distribution of termite (Isoptera) species in
665southwestern Kenya in relation to land use and the morphology of their
666galleries. Biol. Fertil. Soils 3, 69–73.
667Kuznetsova, A., Brockhoff, P.B., Christensen, R.H.B. (2014). lmerTest: Tests for
668random and fixed effects for linear mixed effect models (lmer objects of
669lme4 package). R package vesion 2.0-6. http://CRAN.R-project.org/
670package=lmerTest
671Lal, R., 1988. Effects of macrofauna on soil properties in tropical ecosystems. Agric.
672Ecosyst. Environ. 24, 101–116.
673Lavelle, P., Begon, M., Fitter, A.H., 1997. Faunal activities and soil processes: adaptive
674strategies that determine ecosystem function. Adv. Ecol. Res. 27, 93–132.
675Lee, K.E.,1985. Earthworms, Their Ecology and Relationships with Soil and Land Use.
676Academic Press, Sydney.
677Lobry de Bruyn, L.A., Conacher, A.J., 1990. The role of termites and ants in soil
678modification: a review. Aust. J. Soil Res. 28, 55–93.
679Lubbers, I.M., van Groenigen, K.J., Fonte, S.J., Six, J., Brussaard, L., et al., 2013.
680Greenhouse-gas emissions from soils increased by earthworms. Nat. Clim.
681Change 1–8.
682Mando, A., van Rheenen, T., 1988. Termites and agricultural production in the Sahel:
683from enemy to friend? Netherlands J. Agric. Sustainability 605, 1–7.
684Mando, A., Brussaard, L., 1999. Contribution of termites to the breakdown of straw
685under Sahelian conditions. Biol. Fertil. Soils 29, 332–334.
686Maniania, N.K., Ekesi, S., Songa, J.M., 2001. Managing termites in maize with the
687Entomopathogenic Q15fungus Metarhizium Anisopliae. Insect Sci. Appl. 2, 1–5.
688Mitchell, J.D., 2002. Termites as pests of crops, forestry, rangeland and structures in
689Southern Africa and their control. Sociobiology 40, 47–69.
690Mora, P., Seugé, C., Chotte, J.L., Rouland, C., 2003. Physico-chemical typology of the
691biogenic structures of termites and earthworms: a comparative analysis. Biol.
692Fertil. Soils 37, 245–249.
693Ouédraogo, E., Mando, A., Brussaard, L., 2004. Soil macrofaunal-mediated organic
694resource disappearance in semi-arid West Africa. Appl. Soil Ecol. 27, 259–267.
695Ouédraogo, E., Mando, A., Brussaard, L., 2006. Soil macrofauna affect crop nitrogen
696and water use efficiencies in semi-arid West Africa. Eur. J. Soil Biol. 42, 275–277.
697Ouédraogo, E., Brussaard, L., Stroosnijder, L., 2007. Soil fauna and organic
698amendment interactions affect soil carbon and crop performance in semi-arid
699West Africa. Biol. Fertil. Soils 44, 343–351.
700Palm, C., Blanco-Canqui, H., DeClerck, F., Gatere, L., Grace, P., 2014. Conservation
701agriculture and ecosystem services: an overview. Agric. Ecosyst. Environ. 187,
70287–105.
703Paul, B.K., Vanlauwe, B., Ayuke, F.A., Gassner, A., Hoogmoed, M., et al., 2013.
704Medium-term impact of tillage and residue management on soil aggregate
705stability, soil carbon and crop productivity. Agric. Ecosyst. Environ. 164, 14–22.
706Pulleman, M.M., Six, J., Uyl, A., Marinissen, J.C.Y., Jongmans, A.G., 2005. Earthworms
707and management affect organic matter incorporation and microaggregate
708formation in agricultural soils. Appl. Soil Ecol. 29, 1–15.
709Pulleman, M.M., Creamer, R., Hamer, U., Helder, J., Pelosi, C., et al., 2012. Soil
710biodiversity, biological indicators and soil ecosystem services-an overview of
711European approaches. Curr. Opin. Environ. Sustainability 4, 529–538.
R Core Team, 2013. R: A language and Environment for Statistical Computing. R
712Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.
713org/.
714Rouland-Lefevre, C., 2011. Termites as pests of agriculture. In: Bignell, D.E., Roisin, Y.,
715Lo, N. (Eds.), Biology of Termites: A Modern Synthesis. Springer.
716Sarr, M., Agbogba, C., Russell-Smith, A., Masse, D., 2001. Effects of soil faunal activity
717and woody shrubs on water infiltration rates in a semi-arid fallow of Senegal.
718Appl. Soil Ecol. 16, 283–290.
719Shuster, W.D., Edwards, C.A., 2003. Interactions between tillage and earthworms in
720agroecosystems. In: El Titi, A. (Ed.), Soil Tillage in Agroecosystems. CRC Press,
721Boca Raton.
722Six, J., Elliott, E., Paustian, K., 2000. Soil macroaggregate turnover and
723microaggregate formation: a mechanism for C sequestration under no-tillage
724agriculture. Soil Biol. Biochem. 32, 2099–2103.
10 B.K. Paul et al. / Agriculture, Ecosystems and Environment xxx (2015) xxx–xxx
G Model
AGEE 4997 1–11
Please cite this article in press as: Paul, B.K., et al., Exclusion of soil macrofauna did not affect soil quality but increased crop yields in a sub-
humid tropical maize-based system. Agric. Ecosyst. Environ. (2015), http://dx.doi.org/10.1016/j.agee.2015.04.001
725 Six, J., Bossuyt, H., Degryze, S., Denef, K., 2004. A history of research on the link
726 between (micro) aggregates, soil biota, and soil organic matter dynamics. Soil
727 Till. Res. 79, 7–31.
728 Swift, M.J., Bignell, D. (2001).Q16 Standard Methods for Assessment of Soil Biodiversity
729 and Land Use Practice. ASB Lecture Note 6B, Bogor, Indonesia.
730 Tian, G., Brussaard, L., Kang, B.T., 1993. Biological effects of plant residues with
731 contrasting chemical compositions under humid tropical conditions:effects on
732 soil fauna. Soil Biol. Biochem. 25, 731–737.
733Van Bavel, C.H.M., 1950. Mean weight-diameter of soil aggregates as a statistical
734index of aggregation. Soil Sci. Soc. Am. J. 14, 20.
735Wood, T.G., Johnson, R.A., Ohlagu, C.E., 1980. Termite damage and crop loss studies
736in Nigeria – a review of termite (Isoptera) damage to maize and estimation of
737damage, loss in yield and termite (Microtermes) abundance at Mokwa. Trop. Pest
738Manage. 26, 241–253.
739Wood, T.G., Cowie, R.H., 1988. Assessment of on-farm losses in cereals in africa due
740to soil insects. Int. J. Trop. Insect Sci. 9, 709–716.
B.K. Paul et al. / Agriculture, Ecosystems and Environment xxx (2015) xxx–xxx 11
G Model
AGEE 4997 1–11
Please cite this article in press as: Paul, B.K., et al., Exclusion of soil macrofauna did not affect soil quality but increased crop yields in a sub-
humid tropical maize-based system. Agric. Ecosyst. Environ. (2015), http://dx.doi.org/10.1016/j.agee.2015.04.001