Journal of Arid Environments (2003) 53: 517–528doi:10.1006/jare.2002.1064

Economic assessment of production of flour fromProsopis alba and P. pallida pods for human food

applications.

Peter Felker%}*, Nora Gradosw, Gaston Cruzw & Dante Prokopiukz

%Secretaria de Produccion y Medio Ambiente de la Provincia de Santiago delEstero, 514 La Plata 4200 Santiago del Estero Argentina

wUnidad de Proyectors Ambientales, Universidad de Piura, Piura, PeruzFacultad de Agroindustrias, Universidad Nacional del Nordeste, Roque

Saenz Pena, Chaco, Argentina}Estacion Experimental de Fernandez, Universidad Catolica de Santiago del

Estero, Argentina

(Received 8 October 2001, accepted 11 June 2002)

This paper presents a detailed cost accounting of the processes involved inthe production of flours from the mesocarps of the fruits of Prosopis alba andP. pallida to be used in human foods. The flours from both species are similarand contain 7–8% protein, 0?8–2% fat, 26–32% total dietary fiber and 48–59% sucrose. The data were based on actual production rates in use at theUniversidad de Piura pilot plant using simple machinery. Machinery hourlycharges were based on equivalent rental costs in order to avoid interest chargecalculations. As opposed to other analyses which only included processingcosts delivered to the plant, our analysis ascribed value to the owner of thetrees in order to provide incentives to plant and care for the trees. Thesensitivity analyses showed that the most economically important factor wasthe conversion efficiency of pods into flour. When a 40% conversionefficiency, such as that which has been obtained in both Peru and Argentinawas obtained, and by-product credits for high fiber containing fraction wereassumed, the wholesale selling price with a 5% marketing cost and a 40%profit direct costs was estimated to be $0?66 kg�1 for Peru (using a $0?40 h�1

labor rate) and $1?76 kg�1 for Argentina (using a $1?5 h�1 labor rate).

# 2002 Elsevier Science Ltd.

Keywords: agroforestry; nitrogen-fixing trees; protein; sugar

Introduction

The pods of various species of semi-arid adapted nitrogen-fixing trees of variousProsopis species were an important component in the diets of the indigenous peoples inthe deserts of North America (Felger, 1977), Peru (Cruz, 1988) and Argentina(D’Antoni & Solbrig, 1977). While the pod morphology varied substantially across

*Corresponding author. Fax: +54-8521-2131. E-mail: peter_felker@hotmail.com

0140-1963/03/040517 + 12 $30.00/0 # 2002 Elsevier Science Ltd.

Fig. 1. Partially disected P. alba pod illustrating pod mesocarp, endocarp and seeds.

518 P. FELKER ET AL.

these regions, generally speaking, the high-protein (ca 35%) seeds which constituteabout 10% by weight of the indehiscent pods were discarded due to the hard nature ofthe seed coat (Felger, 1977). Oduol et al. (1986) compared the protein and sugarcontent of North and South American Prosopis species in two contrastingenvironments over 3 years and found high variability among accessions, butremarkable stability in protein and sugar content with respect to environment. Themesocarp, which is the spongy pulp between the pod exterior and the leathery capsule(endocarp) which surrounds the seeds (Fig. 1), was the fraction consumed byindigenous peoples. Probably this fraction was consumed as the sucrose content istypically about 45% (the pods are often 30–35% sugar on whole pod basis) (Grados &Cruz, 1996) and does not contain much fiber. In this paper we use the term ‘Prosopisflour’ to indicate the fine (80–100 mesh) powder obtained from the pod mesocarp byvarious milling and sieving techniques.

While Prosopis flour and wheat flour have approximately equal energy and proteincontents, unlike wheat flour which is taste neutral and which is used for its texturalproperties in stimulating volume increase, Prosopis flour does not have gluten(Saunders et al., 1986) and does not stimulate volume increases. As Prosopis flour has ataste and aroma in the same general class as cafe/cacao/cinnamon/mocha, we believethat Prosopis flour will achieve its greatest economic value in flavoring/aromaapplications rather than any benefit from textural properties.

Over the last 15 years, there have been efforts to bring this flour into wide-scale usein contemporary products through dry-milling techniques to produce flours forpastries, coffee substitutes and by water extraction techniques to produce syrups foruse in flavoring applications (Meyer, 1984; Saunders et al., 1986; Cruz, 1988, 1999;Del Valle et al., 1989; Bravo et al., 1994, 1998; Grados & Cruz, 1996;).

Despite these publications describing products with excellent sensory properties,Prosopis flour is not used in any commercial application at significant scale. In order togauge the market possibilities for similar products, it is essential to have detailed costinformation. The goals of this publication are (1) to assess costs for all aspects of flourproduction (returns to farmer, harvester, processor, profit, etc.), (2) to identify thefactors that have the greatest influence on costs, so that future research anddevelopment may decrease these costs and (3) to identify similarly priced products forwhich Prosopis flour may be competitive for human food applications in Argentina and

FLOUR FROM PROSOPIS ALBA AND P. PALLIDA 519

Peru. This study examined P. pallida which is the native species to Peru and P. albawhich is the most important Prosopis species for human food uses in Argentina(D’Antoni & Solbrig, 1977).

A previous economic analysis of Prosopis flour production from various species(Saunders et al., 1986) assumed a hypothetical and considerably more complicatedprocessing system capable of producing 1000 kg h�1. In contrast, this study usedactual production data from a much simpler set of machinery in the Universidad dePiura (UDEP) pilot plant.

Materials and methods

A partially dissected P. alba pod is presented in Fig. 1 illustrating the location of seeds,leathery endocarp and pod mesocarp. The mesocarp is the economically usefulproduct that is converted to a light-yellow, fine (80–100 mesh), 8–9% protein, 40–55% sugar flour. The endocarp which does not possess flavor, sugar components andvery little protein may be useful due to its high dietary fiber content (Saunders et al.,1986). We define the pod flour recovery factor as the mass of flour produced dividedby the mass of pods input to the process. As the mass of the seeds and leatheryendocarp constitutes about 40% of the pods, a recovery factor of 0?60 wouldconstitute the theoretical maximal recovery.

The economic analysis was conducted using an Excel spreadsheet (which may beobtained from the authors) in which the labor costs, production rate for each processand recovery factors may be independently specified. The advantage of including aseparate line for labor charges for each operation is that it permits the flexibility to usedifferent skill levels and it simplifies the separation of hours per operation from laborrate per operation. The time required for each step and yield for each step werederived from values obtained in the UDEP pilot plant.

It is the local custom in both Argentina and Peru that the pod costs are normallytaken to be the values delivered to the processing plant from wild trees. However, it isimportant to separate this cost into collection costs and returns to the tree owner (evenif he does not collect the pods) in order to stimulate good stewardship of the trees.

As the mature pods fall from the tree at maturity over a 6-week period, the harvestingis done by collecting the pods from the ground at weekly intervals to avoid seed weevil(Coleoptera: Bruchidae) damage, damage from rain and soil contamination.

The hourly machinery charges assumed a monthly rental equivalent to 1/60 of theinitial purchase price, use only for 3 months per year and two daily shifts for an 80-hworkweek. Use of equivalent rental charges avoided the need to involve interest ratecalculations. The packaging costs assumed manually placing the flour in 20-lhermetically sealable plastic buckets. The marketing and profit charges (which caneasily be varied) assumed 5% and 40% of total direct costs that are within the range ofindustry standards.

By-product credits assumed the endocarp fraction was used as a high dietary fiberadditive with a 30% yield from the initial pod weight and a selling price of $100 ton�1

as might be obtained in large-scale industrial applications. Credits resulting from podsrejected for human-use applications assumed the pods were sold for livestock feed at aprice similar to oats, barely or medium quality hay at $100 ton�1. This is acompromise between the prices specified for pod purchase in Peru ($66 ton�1) andArgentina ($125 ton�1).

Due to the high sugar content of the pods, it is necessary to dry the pods to 6%moisture (from about 13% moisture at air temperatures of 351C) to prevent the flourfrom adhering to the inside of the milling equipment (Saunders et al., 1986). If 13%moisture pods are attempted to be ground, thick layers of sticky flour build up insidethe mill and eventually stop the mill (Grados & Cruz, 1996).

520 P. FELKER ET AL.

Grados & Cruz, (1996) have found that two additional steps were necessary thatwere not described in earlier works. First it is necessary to manually sort the pods toeliminate those that were heavily damaged by seed-eating weevils. This damage caneasily be assessed by the presence of insect emergence holes in the pods. Second, it hasbeen found necessary to wash the pods in clean water to eliminate soil that may haveadhered to the pods from lying on the ground.

One simple method to separate the endocarp fraction from the flour fraction is togrind the dried pods in a hammermill with a screen with 8 mm diameter holes. The highspeed of the hammermill will disintegrate the low moisture containing mesocarp into afine flour and the entire endocarp will pass through the holes without being destroyed. Acoarse sieve (18 mesh) can be used to separate the mixture of large endocarp containingseed portions from the finer flour. The portion that passes the 18-mesh screen can bereground to pass an 80-mesh screen to make the fine mesocarp flour. The coarse (418mesh) high-fiber endocarp portion with the seed can be ground for livestock feed or forhigh-fiber human food applications. However, it is worth noting that the greatest insectinfestation occurs by bruchids in the seeds and if the endocarps containing seeds (andbruchids) are diverted from human food uses, the insect contamination for human fooduse applications can be greatly reduced. Two contrasting labor charges of $1?5 and$0?40 h�1 were used for Argentina and Peru respectively.

Mature pods for chemical analyses were collected near the city of Roque Saenz Pena inthe Province of the Chaco in Argentina (for P. alba) and from the experimentalplantations of the Universidad de Piura (for P. pallida). All of the further processing andanalysis was conducted at the UDEP pilot-scale processing facility and chemicallaboratories. The fruits were brushed to remove extraneous material, dried at 601C for60 h in a convection oven which reduced the moisture from 12% to 6% and then milledin a custom-designed mill at UDEP. The chemical analyses were performed on the fineflour fraction that passed a 0?15 mm screen. The moisture content was determined bydrying to constant weight at 701C (Association of Official Analytical Chemists, 1997).The protein was determined by Kjeldahl (N� 6?25) (AOAC 960?52). The fat wasdetermined by AOCS method Aa 4-38 (AOAC 925?10: American Oil Chemists Society,1989). The ash was determined as the residue remaining after heating at 5501C in amuffle furnace (AOAC 923?03). The minerals were measured by a Perkin-Elmer 1100spectrophotometer after ashing at 6001C for 5 h (AOAC 944?02, 944?03). Crude fiberwas measured by AOACS method Ba 6-61. Reducing sugars were measuredvolumetrically using the method of Fehling–Causse–Bonans (Montes, 1981) and totalsugars were measured by the same method after acid hydrolysis (Montes, 1981).Condensed tannins were measured by AOAC methods 14048 and 14049. Total solublepolyphenols were measured spectrophotometrically at 765 nm using the Folin–Ciocalteau reagent (Montreau, 1972). Total carbohydrates were calculated as the totalmass of the sample minus water, ash, protein and lipids (Codigo Alimentario Argentino,1998) and the energetic value was calculated by multiplying the respective fractions by4 kcal g�1 for protein, 4 kcal g�1 for carbohydrates and 9 kcal g�1 for fats (CodigoAlimentario Argentino, 1998). The soluble fiber was the fraction that precipitated withethanol after previous digestions with a-amylase and protease to enzymaticallyhydrolysed starches and proteins while the total dietary fiber was determined on theethanol/acetone-washed supernatant after ash and protein determination by AOACmethods 920?86 and 985?29. The flour color was determined using a Minolta CR-300colorimeter using International Color Commission color coordinates.

Results

A comparison of Prosopis pod flour production using Peruvian and Argentine laborrates and a flour/pod recovery factor of 0?4 and the most probable processing rates is

Table 1. Itemization of costs for processes involved P. alba and P. pallida flours forhuman food uses

Peru Argentina

Raw material return to farmer ($ kg�1) 0?030 0?030Collection rate (kg h�1) 12?000 12?000Labor cost ($ h�1) 0?400 1?500Collection cost ($kg�1) 0?033 0?125

Sort pods (kg h�1) 40?000 40?000Labor cost ($ h�1) 0?400 1?500Sorting cost ($ kg�1) 0?010 0?038Yield fraction (kg good pods) (kg bad pods)�1 0?600 0?600Cost of sorted good pods 0?122 0?321

Wash pods (kg h�1) 100?000 100?000Labor cost 0?400 1?500Washing cost ($ kg�1) 0?004 0?015

Drying rate (500 kg 4 h�1) 125?000 125?000propane cost ($ h�1 = $ kg h�1 and $ 0?4 kg�1) 1?600 1?600Drying machinery purchase price 4000?000 4000?000Cost per hour machinery use 0?625 0?625Labor load and unload dryer (h) 4?000 4?000Labor cost ($ h�1) 0?400 1?500Labor cost 1?600 6?000Fraction loss in weight 0?070 0?070Drying cost of 13% moisture pods 0?031 0?066Drying cost of 6% moisture pods 0?033 0?071

Milling rate (kg h�1) 40?000 40?000Labor rate ($ h�1) 0?400 1?500Machine purchase price 2500?000 2500?000Machine hourly charge 0?391 0?391Milling cost ($ kg�1) 0?020 0?047

Cost of milled flour ($ kg�1) 0?179 0?454

Sieving rate (kg h�1) 80?000 80?000Labor rate ($ h�1) 0?400 1?500Sieve purchase price 3500?000 3500?000Machine hourly charge 0?547 0?547Sieving cost ($kg kg�1) of input flour 0?012 0?026Cost of milled and sieved pods ($ kg�1) 0?191 0?479Recovery factor (kg flour) (kg pods)�1 0?400 0?400Cost of flour ($ kg�1) 0?477 1?199

By-product recovery ($ kg)�1 pods) 0?300 0?300Value ($ kg�1) of by-product 0?100 0?100By-product credit ($ kg�1 of pods) 0?030 0?30

Credit for bad podsFraction bad pods (1�good pods) 0?400 0?400Selling price bad pods ($ kg�1) 0?100 0?100Bad pod credit (kg input) 0?040 0?040Packing rate (kg h�1) 40?000 40?000

FLOUR FROM PROSOPIS ALBA AND P. PALLIDA 521

Peru Argentina

Labor costs 0?400 1?500Packing cost ($ kg�1) 0?010 0?038Material costs ($ kg�1) 0?030 0?030Total packaging costs 0?040 0?068

Total direct costs 0?517 1?266Costs including by-product credits 0?487 1?236Costs including bad pod credits 0?447 1?196Marketing costs (5% of total direct costs) 0?024 0?062Profit (40% of marketing+direct) 0?188 0?503Selling price with by-product credit ($ kg�1) 0?660 1?761

Table 1. Continued.

Fig. 2. Sensitivity analyses of the effect of mesocarp recovery on total cost of flour.

522 P. FELKER ET AL.

presented in Table 1. As expected, the differences in labor rates greatly influenced thefinal price of the Argentine and Peruvian flours of $0?66 and $1?76 kg�1. It issomewhat surprising that the price of the flour changed very little from either high-fiber, by-product credits or from the sale of insect-damaged pods for livestock food.

A sensitivity analysis in Fig. 2 illustrates the very dramatic effect of the recoveryfactor (kg flour output per kg input of pods) on the final selling price for both Peruvianand Argentine labor rates. In the Argentine case, change of the recovery factor from0?10 to 0?40 changed the price from $4?75 to $1?81 kg�1. Recent trials in bothArgentina and Peru have obtained 40–54% yields of flour from pods making thisselling price possible. As discussed later this price would put Prosopis pod flour in thesame price range as other competing products.

The present analysis ascribes little value to the owner of the trees. Obviously, if thetrees with high-quality pods are to be available, it is important to ensure a reasonable

Fig. 3. Sensitivity analyses of pod prices paid to landowner to final flour price.

FLOUR FROM PROSOPIS ALBA AND P. PALLIDA 523

return to the landowner. The sensitivity analysis in Fig. 3 illustrates that there is verylittle influence of initial price of pods on the final flour price and that it should bepossible to increase the return to the landowner.

Further, it would be possible for the landowner to achieve additional revenue if she/he collected the pods. If the landowner managed the trees to provide excellent qualitypods that did not require sorting or a 40% rejection rate, in Argentina his/her share ofthe final pod-selling price could be increased to about $0?32 kg�1.

The percentage of total costs to final selling price is presented in Fig. 4 andillustrates that by far the largest cost is due to the low recovery of flour from the totalmilled pods. The next most important costs in final selling price were profit, the loss insorting good pods from bad pods and the collection costs. Even with the higher laborrates associated with the Argentine conditions, the costs of milling, drying and sortingwere relatively insignificant.

The chemical composition of P. pallida and alba flours made from the mesocarpwith a 40% recovery presented in Table 2 must be viewed with caution as thepods were collected from wild unmanaged trees of unselected genetic stockfrom completely different environmental and edaphic conditions. While the protein,ash and energy values were remarkably similar for both species, the P. alba flourhad a tendency to have higher total sugars but lower dietary fiber than theP. pallida flour. The high concentration of sugar in the flour (48–59%) is remarkable,and as has been confirmed by other authors, the bulk of the sugar is thenon-reducing sugar, sucrose (Saunders et al., 1986; Grasos & Cruz, 1996).The presence of 26% to 32% total dietary fiber in P. alba and P. pallidarespectively would be an important asset in maintaining gastrointestinal tractfunctions. The calcium and iron levels were slightly higher in P. alba than P. pallidaand perhaps could be a reflection of growth on more developed soils from themuch higher rainfall region in Argentina. The color of P. pallida is a slightlymore intense yellow color as indicated by the quantitative measurements. In spite of

Fig. 4. Relative importance of costs and processes associated with production of Prosopis flour.

Table 2. Comparison of the nutritional properties and color of P. alba andP. pallida flours derived from pod mesocarp on a dry matter basis

Prosopis alba (g kg�1) Prospis pallida (g kg�1)

Moisture 25?771?2 33?971?0Protein 71?770?2 81?170?8Fat 21?771?0 7?771?2Ash 31?370?2 36?071?9Calcium 1?2770?01 0?7670?04Iron 0?4570?01 0?3370?03Crude fiber 24?370?3 34?070?5Total carbohydrates 849 826Total sugars 59170?9 48572?5Reducing sugars 27?671?0 21?470?8Condensed tannins 5?770?4 4?170?3Soluble polyphenols 0?0670?9 0?1371?0Insoluble dietary fiber 200?9 306?0Soluble dietary fiber 64?7 16?2Total dietary fiber 265?6 322?2Energy value (kcal kg�1) 3780 3620

Color, co-ordinatesL 76?5770?06 65?3970?02a 1?9570?03 6?2470?16b 18?3770?13 21?7270?08

Note: Mean values7standard deviation (n = 3). Standard deviations could not becalculated for the total carbohydrates and energy values as they are sums of othercomponents.

524 P. FELKER ET AL.

the similarities in the chemical composition, the tastes, aromas and color aresignificantly different and it is easy to distinguish the Peruvian from the ArgentineProsopis flour.

FLOUR FROM PROSOPIS ALBA AND P. PALLIDA 525

Discussion

When the highest conversion factor from pods to flour is used, the flour from bothArgentina and Peru was less than $2 kg�1. In contrast, when low conversion rates wereused, the price rose close to $5 kg�1. Saunders et al. (1986) reported a processing costof $0?47 kg�1 of input pods when 576 ton was processed annually, but this decreasedto $0?23 kg�1 with an annual processing of 2300 ton per season. As Saunders et al.(1986) found, the fine flour fraction to vary from 36% to 46% for the entire P. velutinapod, the cost of the flours from their processing would range from $1?3 to 1?02 kg�1

(assuming no by-product credits). However this does not include the initial purchaseprice of the pods. Del Valle et al. (1989) developed a simple method for producingProsopis flour for human food applications which consisted of toasting, pin milling andscreening the flour through 45- and 100-mesh screens. These authors did notfractionate the pods into various components and all of the pod material that passed a100-mesh screen was deemed ‘Prosopis flour’. Thus the Del Valle et al. (1989) flourwas lower in sugar 28% vs. 46% but higher in protein 10?1 vs. 8?1 than the Peruvianflour (Grados & Cruz, 1996) that was made from the mesocarp. These authorsconducted taste panel evaluation with nine food products ranging from mixtures withchocolate beverages to rolled oat mixtures and found that the chocolate/Prosopisbeverage and a pressed Prosopis/peanut butter candy had the greatest acceptability.The processing cost of the flour was determined to be about $0?66 kg�1 but again thisdid not include the cost of producing or collecting the pods.

In an economic analysis of Prosopis pod processing for livestock feed by a co-operative in the Central Highlands of Mexico, Silbert (1996) reported that in 1987Mexican farmers were paid $0?127 kg�1 by the co-operative but that other ranchersbought Prosopis pods from non-co-operative sources for $0?018–0?095 kg�1. Thesepods were simply ground by a hammermill without drying but there were considerableproblems with clogging of the grinders due to high moisture content of the pods.Silbert (1996) estimated that the cost of ground, unselected pods was $0?079 kg�1

using a 10% interest rate, a 2000 ton per year production capacity and a purchaseprice of $0?063 kg�1. Even for livestock feed uses, we believe this selling price wouldnot be sustainable because it does not include a profit to the landowner, costs tomaintain the trees and a pod-drying procedure which is essential in maintainingconsistent output.

As Prosopis flour does not have gluten (Saunders et al., 1986), it cannot besubstituted for wheat flour in applications where increase in volume during baking isimportant. However gluten-free flour would be useful in applications for people withgluten intolerance. For example, Cruz (1988) found that as little as 5% P. juliflora floursignificantly decreased the dough expansion of 100% wheat flour bread but theresulting bread had excellent appearance, a good smell and was less elastic than 100%flour bread. Cruz (1988) evaluated biscuits (use of sodium bicarbonate instead ofyeast to increase volume) made from 5–24% Prosopis flour to overcome the problem ofvolume decrease in yeast-type breads. The general consensus of clients of theUniversidad de Piura cafeteria, where the Prosopis biscuits are sold as a routine item,was that the 14% and 24% flour-containing biscuits had superior quality and taste tonormal biscuits. Similar findings were observed by Saunders et al. (1986) who foundthat Prosopis flour could constitute up to 10% in leavened bread, up to 50% inchapatis, crackers and flakes and up to 20% in tortilla corn chips.

Taste panels in trials of Cruz (1988) noted a bitter aftertaste in some breadmade with Prosopis. Pods of some Prosopis alba in Argentine progeny trials also werenoted to be very astringent or bitter (Felker et al., 2001) and work is in progressto identify the compound responsible for this taste. Preliminary indications arethat it is a saponin (Fabiani pers. comm., 2000). Genetic improvement trials in

526 P. FELKER ET AL.

both Peru and Argentina (Felker et al., 2001; Alban et al., 2001) have ranked trees asvery bitter, bitter, sweet or very sweet and only cloned trees with pods that were sweetor very sweet.

The percentage of protein in the pods was much lower than that of the seeds.However when care was taken to avoid the destructive effects of carbohydrates onsulfur amino acids and tryptophan during acid hydrolysis, Felker &Bandurski (1977)found that the amino acids which are normally most limiting in legumes were higher inthe pod mesocarp than in the seeds, e.g. 76 vs. 57 g g N�1 for methionine, 38 vs.14 mg g N�1 for cysteine and 112 vs. 61 mg g N�1 for tryptophan. As there wassignificant non-dialysable, non-trichloracetic acid precipitable N in the flour ofProsopis pods, it was hypothesized that the pods contained glycoproteins (Felker &Bandurski, 1977). For a dried flour, the Peruvian Prosopis has moderately highconcentrations of vitamin C and nicotinic acid, i.e. 60 and 3 1 mg kg�1 respectively(Grados & Cruz, 1996).

It is our opinion that Prosopis flour will likely be used in applications where its cafe/mocha/cinnamon-like aroma and taste will be used for flavor enhancement rather thanits structural/rheology properties. The chemical composition of the flavor and aromahas not been reported. It is our opinion that the aroma and taste fall in the samegeneral category as sweet (as opposed to hot pepper) aromatic spices such as nutmeg,cinnamon, cacao and cloves. This aroma and flavor is concentrated in the mesocarpfraction. We feel competing products will be cacao or carob flour. Informal sensorialanalysis by industry and university personnel in Peru and Argentina suggests that theoptimum concentration of Prosopis flour in cookie-type pastries and granola bars is inthe 12–18% range. Thus at a cost of $1?80 kg�1 of flour to the wholesale buyer, thecost of the Prosopis portion would range from $0?21 to $0.32 kg�1 of the finalproducts. As granola-type products and cereals have a retail price close to $10 kg�1,the Prosopis portion would constitute less than 10% of final selling price.

According to the World Bank commodity report (www.worldbank.org/html/ieccp/pkjan98.html) the average world prices for cocoa and mild coffee were $1.61 and$4.16 kg�1 in 1997 and thus would be comparable to our wholesale prices for Prosopisflour. The retail price of carob (Ceratonia siliqua) flour sold over the internet for thegluten-free celiac market was $8.66 kg�1 (www.glutenfreemall.com). As Prosopis flouralso does not contain gluten, it would appear that celiac market would be an excellentinitial route to introduce Prosopis flour to the commercial market.

From an agronomic perspective, Silbert (1996) has suggested that in Mexico thehigh year-to-year variability in production (with some years having very little podproduction) will be one of the major constraints to pod commercialization. High year-to-year variability may also occur in Argentina and Peru. However if ground flour isprotected from insect attack and humidity, flour from good pod harvests could bestored for several years until a good crop became available. While there has been muchspeculation about factors such as heavy rains and winds during pollination that mightaffect pod production, there is surprisingly little quantitative data on this subject.However, Lee & Felker (1992) reported that pod production over five rainfed sites was3.3 times greater in a drought year as compared to a normal year and that beevisitation was not associated with nectar production but was associated with podproduction.

Currently in both Argentina and Peru, low pod prices of about $30–60 ton�1 arepaid at the processing plant. This obviously only represents the labor cost to harvestthe pods. At this price there will be no incentive to plant and care for more trees andthe trees that exist have no value to the landowner and may as well be harvested forcharcoal or lumber. If the owner could achieve a return of $30 ton�1 independent ofcollection costs, in good years with a high yield of 3 ton ha�l from a well-managedplantation or native forest, he would achieve a revenue of $90 ha�l and in many semi-arid regions this would be a very significant income.

FLOUR FROM PROSOPIS ALBA AND P. PALLIDA 527

At present, all of the pods used for flour in both Argentina and Peru are collectedfrom wild stands of native unselected genetic stock. Due to great tree-to-treevariability in pod production and pod flavor, genetic improvement programs in Peruand Argentina have identified and cloned trees with superior height growth, podproduction and pod flavor (Felker et al., 2001; Alban et al., 2001). As Wojtusik et al.(1993) found that Peruvian Prosopis in a Haitian progeny trial was the fastest growingand most erect accession in the trial, they also made clones of the Peruvian Prosopis.Due to the very extensive areas in which wild unselected Prosopis occurs in Peru,Grados & Cruz (1996) have suggested the total resource potential of the PeruvianProsopis is about 300,000 tons year�1 The magnitude of the Prosopis pod resources inArgentina is uncertain but as Prosopis occurs widely in five Argentine Provinces ofSantiago de1 Estero, Salta, Catamarca, Chaco and Cordoba, the Argentine resourcepotential is probably as great as in Peru. We believe contemporary processingtechniques, when combined with sound economic analyses and aggressive marketinghave the capability to radically improve the economic well-being of the resource-poorinhabitants of the arid regions of both Argentina and Peru.

References

Alban, L., Matorel, M., Romero, J., Grados, N., Cruz, G. & Felker, P. (2001). Cloning of elite,multipurpose trees of the Prosopis juliflora/pallida complex in Piura, Peru. AgroforestrySystems, 54: 173–182.

American Oil Chemists Society. (1989). Official and Tentative Methods, Sampling and Analysis ofOilseed by Products, and Sampling and Analysis of Vegetable Oil Source Materials. USA.

Association of Official Analytical Chemists. (1997). Official Method of Analysis of AOACInternational, Vol I and II (16th Edn). Arlington, VA, U.S.A.:AOAC.

Bravo, L., Grados, N. & Saura-Calixto, F. (1994). Composition and potential uses of mesquitepods (Prosopis pallida L): comparison with carob pods (Ceratonia siliqua L). Journal of ScienceFood and Agriculture, 65: 303–306.

Bravo, L., Grados, N. & Saura-Calixto, F. (1998). Characterization of syrup and dietary fibreobtained from mesquite pods (Prosopis pallida L). Journal of Agricultural and Food Chemistry,46: 1727–1733.

Codigo Alimentario Argentino. (1998). Marzocchi Ediciones, Buenos Aires.Cruz, G.E. (1988). Evaluation of flour from Prosopis juliflora and Prosopis pallida pods in bakery

and extrusion-cooking products. In: Habit, M. (Ed.), The Current State of Knowledge on Prosopisjuliflora, pp. 425–439. Food and Agricultural Organization, Rome Publishers. 554 pp.

Cruz, G. (1999). Production and characterisation of Prosopis seed galactomannan. Ph.D thesisDiss. ETH 13153, Swiss Federal Institute of Technology, Zurich. 118 pp.

D’Antoni, H.L. & Solbrig, O.T. (1977). Algarrobos in South American cultures past andpresent. In: Simpson, B. (Ed.), Mesquite. Its Biology in Two Desert Ecosystems, pp. 189–199.Stroudburg: Dowden Hutchinson and Ross Publ. 250 pp.

Del Valle, F.R., Marco, E., Becker, R. & Saunders, R.M. (1989). Development of productscontaining mesquite (Prosopis spp) pod flour and their nutritional and organoleptic evaluation.Journal of Food Processing and Preservation, 13: 447–456.

Felger, R.S. (1977). Mesquite in Indian cultures of southwestern North America. In: Mesquite.Its Biology in two Desert Ecosystems, Simpson, B.B. Hutchinson and Ross Publ. 250 pp.

Felker, P. & Bandurski, R.B. (1977). Protein and amino acid composition of tree legume seeds.Journal of Science of Food and Agriculture, 28: 791–797.

Felker, P., Lopez, C., Soulier, C., Ochoa, J., Abdala, R. & Ewens, M. (2001). Genetic evaluationof Prosopis alba (algarrobo) in Argentina for cloning elite trees. Agroforestry Systems, 53: 65–76.

Grados, N. & Cruz, G. (1996). New approaches to industrialization of algarrobo (Prosopispallida) pods in Peru. In: Felker, P. & Moss, J. (Eds), ‘Prosopis: semi-arid fuelwood and forage tree.Building Consensus for the Disenfranchised’, U.S. National Academy of Sciences, March 1996.Kingsville, TX: Center Semi-Arid Forest Resources Publ. www.tamuk.edu/webuser/symno-sium. 1?1–6?60.

528 P. FELKER ET AL.

Lee, S.G. & Felker, P. (1992). Influence of water/heat stress on flowering and fruit of mesquite(Prosopis glandulosa var. glandulosa). Journal of Arid Environments, 23: 309–319.

Meyer, D. (1984). Processing, utilization and economics of mesquite pods as a raw material forthe food industry. Ph.D. thesis Diss. ETH 7688, Swiss Federal Institute of Technology. 159pp.

Montes, A. (1981). Bromatologıa. Editorial Universitaria Buenos Aires, II: 315–316.Montreau, F.R. (1972). Sur le dosage des composes phenoliques totaux dans les vins par la

methode Folin-Ciocalteau, Journal International des Sciences de la Vigne et du Vin, 24: 397–404.Oduol, P.A., Felker, P., McKinley, C.R. & Meier, C.E. (1986) Variation among selected Prosopis

families for pod sugar and pod protein contents. Forest Ecology and Management, 16:423–431Saunders, R.M., Becker, R., Meyer, D., Del Valle, F.R., Bellestin, E.M. & Torres, M.E. (1986).

Identification of commercial milling techniques to produce high sugar, high fiber, high proteinand high galactomannan gum fractions from Prosopis pods. Forest Ecology and Management,16: 169–181.

Silbert, M. (1996). A mesquite pod industry in Central Mexico: an economic developmentalternative. In: Prosopis: Semi-arid Fuelwood and Forage Tree. Building Consensus for the Disen-franchised, US. National Academy of Sciences, March 1996. Felker, P. & Moss, J. (Eds),Kingsville, TX: Center Semi-Arid Forest Resources Publ. www.tamuk.edu/webuser/sympo-sium. 1?1–6?60.

Wojtusik, T., Felker, P., Russell, E.J. & Benge, M.D. (1993). Cloning of erect, thornless, non-browsed nitrogen fixing trees of Haiti’s principal fuelwood species (Prosopis juliflora).Agroforestry Systems, 21: 293–300.