·!" /b.b14 Efforts to develop cassava technology for the vast areas of acid infertile soils continued along the strategy outlined last year (CIAT Cassava Prog: 1979 Ann. Rept.). The screening of cassava germplasm in nutrient solution was discontinued, partly because of difficulties in reproducing results dueto plant variability, and partly because nutrient solution screenings do not account for the differential ability of varieties to forrn mycorrhizal associations that are so essential for P uptake. Large-scale field screenings were begun in Carimagua for P and acidity tolerance, and in CIA T-Quilichao for P tolerance. The second aspect of improving the efficiency of nutrient absorption and fertilizer application involved various fertilizer trials to determine: a) nutrientabsorption and distribution withín the plant during a J2-month growth cycle; b) the long-term effect of N-P-K applications on soil fertility and yíeld of continuously grown cassava; e) the residual effect of varíous sources of P; d) the lime x P interaction; and e) the effect of mycorrhizal inoculation on P absorption. Screening for Low-P Tolerance A small-scale screening for P tolerance in CIAT- Quilichao continued. Thirty-two cultivars from the germplasm collection were planted wíthout P and with 44 kg P 1 ha ( 100 kg P2 Os) applied as triple superphosphate in plots that previously had received 88 and 44 kg P 1 ha in two preceeding years, respectively. Leaf samples were taken at three months. At 12 months plants were harvested. Table 1 shows P content of leaves, fresh root yields, starch content, P tolerance index and percentage Table l . Effect of P fen iJization on leaf P content, root yield, starch content and mycorrhizal infection ofthe 10 most low-P -tolerant cassava cuJtivars, a t CJA T-Quilichao. Cultivar P content in YFEL 1 (%) Root yield 2 (t / ha) Root starch (%) P applied (kg/ ha) o 44 o 44 o 44 ICA-HMC-2 0.36 0.35 37 33 22 19 M Col 1226 0.39 48 56 26 24 M Mex 59 0.40 0.38 44 49 28 29 M Col 1879 0.28 0.32 37 36 26 27 M Col 1684 0.31 0.29 36 39 25 25 M Col 113 0.30 0.40 42 58 26 25 M Col 131 0.26 0.28 31 32 M Ven 83 0.30 0.35 36 49 30 31 M Col 88 0.33 0.41 32 39 Llanera 0.36 0.42 24 23 1 YFEL • youngest fu lly e~panded lea ves at three months after planting. 2 Single row yield, average of thrce plants, four replications. l % of total root observations having hyphae, vesicles or arbuscles. • P-tolerance index: Yicld (O P) Yield (0 P) X 100 X Yield (44 P) Max. Yield (0 P) MycorrhizaJ infection 1 p. (%) tolerance index • o 44 45 44 86 . 25 56 85 35 34 83 37 30 80 34 24 70 39 37 63 40 36 62 40 26 57 22 39 55 43 50 54 59 mycorrhizal infection of JO of the most P-tolerant cultivars. The P tolerance index of the 32 cultivars ranged f ro m 11 to H6 with an average of 43. Leaf P content at three months and the varieties' P tolerance index were not related as P content reflects both P uptake ability and plant size, which is highly variable among cultivars. Average root yields of the ten most tolerant cultivars increased only 10% by the application of P. Starch content increased ip sorne cultivars and decreased in others by P fertilization, but dífferences were not significant at the 5% leve! for any of the seven cuJtivars studied; starch content varied significantly among cultivars. Mycorrhizal infections, especially vesicle numbers, were higher in P-fertilized plants than in those witbout P. However, there was no correlation between percentage infection of roots at harvest, and root yield or P tolerance index. Changes in root infection during the growth cycle will have to be determined in order to establish whether early infection is essential for obtaining high yields in low-P soils. Nutrient Absorption and Distribution To determine the effect of fertilizers on the rate of nutrient absorption during dífferent stages of plant development and nutrient distribution within the plant (nutrient profile), two cassava cultivars were planted in CIAT-Quilichao in large plots with and without applied fertilizer. Cultivars were M Col 22 and M Mex 59, the former a non-vigorous plant type and the latter a very vigorous one. All plots were limed with 500 kg{ha of dolomitic lime and the fertilized plots received 1 t f ha 10- 30-10, 20 kg S/ ha as elemental sulphur, JO kg Zn fha as ZnS04.1H20, and 1 kg Bf ha as Borax, all applied broadcast and incorporated before planting; another 50 kg N f ha as urea was applied at 60 days. Two plan_ts per plot were harvested monthly and separated into upper, middle, and lower leaf blades ("lea ves"), petioles and stem, as well as roots. Samples were dried, weighed and analyzed for a11 macro- and micronutrients. Soil samples were also ta.ken monthly and analyzed. Figure 1 shows the rate of total dry matter (DM) production for the two varieties. Fertilizer had no significant effect on M M ex 59 (a variety highly tolerant to low P and not responsive to high levels of fertility- ClA T Ann. Rept. 1977, 1978and CIATCassava Progr.l979 Ann. Rept.). M Col22, on the other hand, produced significan ti y more DM when fertilized, especially during tbe last six months. Effects weresimi.larfordryrootyields (Fig. 2). In M Mex 59 root yields were highly variable during the last 60 ... 1 three months with no apparent response to fertilization, while M Col 22 yielded similarly without fertilizer, but produced much higher yields with fertilization. In both varieties fertilization stimuJated top growth more than root growth, decreasing harvest index about 10- 15%. Unfertilized M Col 22 had a final harvest index of about 0.70..0.75, which remained constant during the last five months, while the harvest index of M Mex 59 continued to increase until the 10th month, but reached only 0.55. Thus, the more vigorous M Mex 59 produced relatively more to growth, while the nonvigorous M Col22 produced more roots and atan earlier stage. At six months, M Col22 had produced about 50%, while M Mex 59 had produced only about 25% of its final root yield. ~ ;: .i1 c. 00 ~ ] o ::l "8 2.0 a. .. ~ ~ E ;:.., o 1.6 1.2 0.8 0.4 o o Figure l. 4 Montlls O Fertilized e Nonfert1hzed Cumulati•·~ total dry matter production ov~r 12 monthsfor two ca.ssava l'ultivar.s grown with and wuhout fertiliz~rs, at CIA T-Quilichao. • ~ e: 01 o. ~ "O 8 :S "O o .... Q. .... ~ ;:; E ;>, .... "O o o a: Figure 2. 0.6 0.4 0.2 o 1.0 0.8 Months O Fertilized e Nonfertilized Cumulatil't~ root dry matter production over 12 m onthsfor two cassava cu/11vars grown with and withoutfertilizers, at C/A T-Quiltchao. Figure 3 shows the distribution of DM between roots , stems, leaves and petioles during the cultivars' growth cycles without fertilization. Roots started to accumulate DM after two months, and reached a maximum at JO months in both cultivars; however, these cultivars varied markedly in their DM distribution with M Col 22 translocating most D M to roots after t he third month while M Mex 59 did so only after the eighth month. The DM distribution in fertilized plants followed essentially the same pattern, except that in M Col 22 roots continued to accumulate DM up to the 12th month. 1 n general, N, P and K contens of the various plant parts decreased with time, especially in the upper stem tissue. Nutrient contents of upper lea ves also decreased with time but not to the same extent as those of stems, petioles or roots, making this tissue more suitable for diagnostic purposes. For most elements, sampling at three months is recommended , while ata la ter date of sampling the critica! levels should be reduced. The contents of most other nutrients also decreased with time, except Fe and Mn which remained constant in upper leaves, and Ca which increased in upper lea ves and in petioles. Nutrient contents of roots decreased markedly for all elements as starch increased with time. lt is clear that roots are relatively high in N and K, and these elements are removed in greatest quantities in each root harvest. The nutrient contents of all plant parts at 2-4 months of age are shown in Table 2, both for fertilized and non- fertilized plants. Fertilization had mainly increased the nutrient content of all tissues without significantly changing the distribution pattern. Nitrogen and P contents followed similar distributional patterns and decreased in all plant parts from the top to the bottom of the plant. Potassium was highest in upper stems followed by petioles and leaves, however, the K gradient from upper to lower stem was much greater than for leaves. Leaves were more indicative of K status than the upper stem or petioles; the lower stem or the petioles might also be good indicator tissue for K although the former would destroy the plant during sampling. Data for oth~r nutrients CI able 2) indica te Ca and Mg were both about equally high in lea ves, petioles and stems, and low in roots. Unlike N, P and K, lower leaves and petioles were higher in Ca and Mg than upper leaves. Sulphur was high in lea ves, extremely low in petioles and intermediate in stems; 8 was rather uniform throughout the plant; Cu was high in stems; Fe was high in leaves, especia\ly lower leaves; while Mn and Zn were high in petioles especially lower petioles. The Fe, Mn and Zn contents all increased from upper to lower leaves and petioles. In this extremely acid (pH 3.9-4.1) soil, tissue concentrations of Fe and Mn were very high, probably above the critica! leve! for toxicity. Figure 4 shows the cumulative N, P, K profiles in M Col 22. Plants continued to accumulate nutrients throughout the growth cycle, but the greatest rate of absorption occurred from the second to the ftfth month, especially for K. After the fifth month all these elements had ac- cumulated mainly in the roots of this early root bulking cultivar. At harvest, quantities of N, P and K weregreatest in roots (66% for K) followed by stems, lea ves and petioles. However, Ca, Mg and Mn accumulated more in the stem than in the roots. 61 900 M Mex 59 800 700 600 ~ e: " c. ::E 500 'O " u "' 'O~ Q. 400 ... ~ ;;; E ;., o 300 o o 2 4 6 8 10 12 12 Months Roots • Lea ves o Stems o Petioles e Figure 3. Distri!Jution of dry matter among roots, stems, leoves ond petioles in two cassavo cultivars grown without fertilizer o~·er 12 monrhs, ot CIA T-Qul/ichoo. Table 2. Concentration of nutrients in cassava upper, middle and lower leaves, petioles and stem, and rootst. Plant par! N utrient content (%) Nutrient content (ppm) N p K Ca Mg S B Cu Fe Mn Zn Le aves upper 5.75 0.42 1.98 0.72 0.34 0.30 11.1 12.2 176 400 107 middle 5. 18 0.27 1.80 1.01 0.38 0.28 11.9 f2.1 237 523 119 lower 4.40 0.20 1.58 1.34 0.49 0.22 11.7 11.1 386 697 137 Petloles upper 2.25 0.22 2.93 0.90 0.38 0.06 10.6 9.0 66 533 90 nidd le 1.41 0.14 2.35 1.13 0.39 0.02 9.9 7. 1 56 !05 127 lower 1.35 0. 12 2.23 1.54 0.48 0.01 10.9 7.5 123 1470 190 Stem upper 2.73 0.30 3. 15 0.82 0.37 0.18 10.5 18.1 133 339 K6 middle 2.21 0.27 2.21 1.02 0.38 0.16 8.6 22.7 107 379 120 lower 1.28 0.22 1.14 0.65 0.31 0.09 6.4 23.6 225 170 97 Roots 1.52 0.18 1.56 0.24 0.14 0.05 6.0 10.7 508 17K 66 A'crages ol samples taken at two. three and four months, for fertilized and unlenihzed culttvars M Col 22 andM Mex 59 62 20 N -E "' Ci. -.. ~ ., u ::> ., -~ p -= .: 0.8 u E o (.) E 0.4 -~ :; ;z o 12 10 8 K 6 Months A Total o Stems • Roots O Leaves e Petioles Figure 4. Total uptake and distribution of N, P and K among plant parts of cassava cultivar M Col22 over 12 months. at CIA T- Quilichao. (A verages from fertilized and nonfertilized plants.) Monthly -analyses of unfertilized soils showed no significant changes in pH and Ca content, a slight increase in exchangeable Al from 2.5 to 3.5 meq/ 100 g; a slight increase in available P from 4 to 6 ppm, possibly dueto leaf fall and decomposition; a slight decrease in exchangeable Mg; and a significant decrease in exchangeable K, from 0.4 to 0.2 meq/ 100 g. In fertilized plots, exchangeable K decreased from 0.60 to 0.25 meq / 100 g, and increased again during the final month. The marked decrease in soil K isthe cause of soil exhaustion after cassava cultivation and the reason why cassava responds more to K fertilization after severa! consecutive cassava crops. lf each plant of a cassava crop removes about 10 g of K (see Fig. 4) and all was absorbed from the top 20 cm of soil without replacement, this decrease is as much as 0.17 meq K/ 100 g of soil; additional K may be lost by leaching and erosion. Thus, adequate K fertilization is essential to obtain high yields while maintaining soil fertility (see next section). Long-Tenn Fertility Trial Objetives, experimental treatments and results of the first planting with cultivar Llanera were described last year (CIAT Cassava Progr. 1979 Ann. Rept.). In the second year, cultivar CMC 40 was planted in the same plofs without additional fertilizer, except the eight additional plots receiving fertilizers annually. Figure 5 shows response to residual effects of N, P and K for root and foliage yield, harvest index, and N, P, K contents of leaves of three-month-old plants. Despite no additional application of fertiliz.ers, root yields were very high with the lowest yield (31 t / ha) obtained with an absolute zero fertilization of N, P and K. In general, plants responded to fertilization more in terms of foliage than root yields but unlike the first year, there was no significant decrease in harvest index due to fertilization. Figure 6 shows the root yield response to zero, medium and high fertilization both for the single and annual applications. Although there was a significant residual effect from the initially. applied fertilizers, root yields increased another 5-6 t/ ha from reapplication of these fertilizers. This would have been economical at the intermediate leve! of application, but not at the highest leve!. After two years of cassava cropping tbe available P content of the soil had decreased from 7.8 to 2.3 ppm without applied P and from 41.8 to 6.9 ppm witb 175 kg P/ ha applied. Similarly, without additional K the 63 • 40 ._ • • ~ ~ .2 1.0 -u 00 >( .!! u o 30 5.0 .4 2.0 "O .... 6 ~ .S ... IJr :.¿ ~ o z c. "' ~ o ~ ~ .75 "' e ~ :e.... 4.0 1.0 o "O -.; ;;: .50 o 100 200 o 87 175 o 125 N (kg/ ha) P (kg/ ha) K (kgf ha) ARoots O f o1iage e Harvest index lJ. N P K contenl ili 1eaves F1gure 5. Residual effecls f rom 1hree levels of N, P and K applied lo a previous cassava crop on nutriem con1em s of upper leaves, al 1hree momhs, and on roo1 and joltage y ields and harvest index of 1 Z-momh-old cultivar CM C 40, at C/A T-Quilichao. 50 exchangeable K content of the soil decreased from 0.20 to 40 -;;- .S:: -- "O -.; 30 ·:;, o o ... .S:: "' ~ 20 1.1.. 10 Figure 6. 64 100-87-125 200-175-250 N-P- K app1ied (kg/ ha) O Residual effect e Annua1 app1ication Root yield response of cultivar C MC 40 to zero, medium and high levels of N, P and K applied only to the preceding crop (residual effect) or when applied annually, at CIA T- Quilichao. 0.12 meqj 100 g, and with the initiaJ application of 250 kg K jha, K content decreased from 0.48 to 0.14 meq/ 100 g. Only with annual reapplication of 250 kg Kj ha could the soil K content be maintained at 0.21 meq j 100 g after the second cropping, while an annual application of about 90- 100 kg P 1 ha would be required to maintain an available P content of 7 ppm in the soil. Sources of P Two experiments to study the effect of various P sources, levels and methods of application on cassava yields in CIAT-Quilichao were initiated in 1978 (CIAT Cassava Prog. 1979 Ann. Rept.), and repeated in 1979 to study the residual effect. In the first year root yields of Llanera cassava varied from 20 to 25 t j ha, without a significant response to P application. In the second year, without additional P applied, roor yields of cultivar M Col 1684 varied from 42 to 51 t / ha, again without a significant response to P. Phosphorus contents of upper lea ves at three months varied from 0.38 to 0.42% in the check plots; this is at the criticallevel of 0.4%, indicating an adequate Pstatus of the plants. In the P check plots the average yield was 46 t fha. This high yield was attained in a soil with an initial P content (Bray ll) of only 3 ppm, which increased to 5-6 ppm due to organic matter mineralization even before the first planting, and had remained at that leve! at the second planting. Plants in all treatments were highly infected (48- 83%) with mycorrhiza, resulting in an efficient uptake of P even from a very low-P soil. Mycorrhizal infection was particularly evident in plots receiving no or only small amounts of soluble P, or high amounts of rather insoluble phosphate sources; infection decreased witb increasing amounts of soluble P applied. Starch content of roots varied from 26 to 28% with P application having no significant effects. These trials suffered a 30-50% defolia- tion at eight montbs due to a severe homworm attack, followed by three months of extreme drought (total of 57 mm rainfall) without any apparent detrimental effect. Lime x P Interaction An experiment has been conducted for two consecutive years in Carimagua to determine the interaction between lime and P applications on an acid infertile soil (CIA T Ann. Rept. 1976). X 16 Jevels of lime 92 kg p o~~~~~·~~~--~~ o 60 120 180 240 300 360 420 480 P2o5 applied (kg/ ha) e Roots Lime applications from zero to 4.8 t / ha increased soil pH from 4.15 to 4.75, decreased exchangeable Al from 3.5 to 1.3 meq¡ 100 g and decreased Al saturation from 88 to 27%. Root yields of M Col638 ranged from 7 t/ ha without lime and P to 25 t/ ha with a combination of 3.6 t / ha of lime and 380 kg¡ ha of P205. Figure 7 shows root yie.ld and foliage growth responses to several rates of P and lime applications. On the average, near maximum root yields wereobtained with 210 kg Pz Os / ha (92 kg P) and 1.1 t lime/ ha. While root yields showed a more or less quadratic response, foliage yields showed a nearly linear response to P, as was also reported in 1976. Harvest index was highest (0.65) with 0.3 t lime and 21 O kg Pz05 and decreased to 0.50 at the highest lime and P rates. In Carimagua, near maximum yields can apparently be obtained with the application of90-IOO kg P / ha and 1 t/ ha of dolomitic lime, which produced a pH of 4.3 with 2.75 meq Al / 100 g or 76% Al saturation. These data indicate that cassava is extremely tolerant of low pH and high Al, but also illustrate the beneficia) effect of small lime applications. o O Fo liage 1.1 t lime .6 1.2 1.8 2.4 3.0 3.6 4.2 Lime appied (t/ ha) 4.8 Figure 7. ::.fject of increasing /evels of applied P and lime on root and fo /iage yields of cultivar M Co/638, in Carimagua. Arrows indica te rates required for 9.5% of maximum yield. 65 Response to Mycorrhizallnoculation In many cassava growing areas of Latin America, P deficiency is the main Jimiting nutritional factor. Inocula- tion with mycorrhizal fungi has been shown to improve cassava's ability to absorb P from low P concentrations in both soils and nutrient solutions (CIAT Cassava Prog. 1979 Ann. Rept. ). Cassava plantlets produced in a misting chamber were planted in pots with sterilized soil from CIAT-Quilichao to which nine levels of P had been applicd. Plantlets were either noninoculated or inoculated with mycorrhizal spores or infected roots. A clear P response could be observed after two weeks anda response to root inoculation after three weeks, the latter becoming more pronounced with time. The dramatic response to inoculation can be seen in Figure !L Figure 44. Response of ca.ssava cultivar M M ex 59 to mycorrhizal inoculation and .severallevei.J of applied P. in a sterilized soil from CIA. T-Quilichao. 66 Without inoculation plants remained P-deficient even with 800 kg P ¡ ha applied, and reached near-maxim u m growth only with 1600 and 3200 kg Pf ha. With root inoe'Ulation, plant growth was very good even without applied P; response to P-application was small initially and had nearly disappeared by harvest. The response to spore inoculation was essentially zero. lnoculation with 2 g of infected ·roots in the P check increased top growth over 80 fold compared with the noninoculated treatment, and was about equally effective as the application of 1600 kg P/ ha (Fig. 9). Without mycorrhizal infection, plants absorbed very little P from a soil to which 800 kg P 1 ha had been applied. In the ClA T-Quilichao soil, mycrorrhizal ínoculation was highly effective even without any P applied, althouth infection and the P concentration of the tops were relatively low a t this leve! (Table 3). Maximum infection occurred at intermedia te P levels corresponding to 50 and ._ o 100 kg P; ha applied (levels used in field-grown cassava). Without P applied, total uptake was o ver 100-fold higher in mycorrhizal than in non-mycorrhizal plants; even at the highest levels of applied P, mycorrhízal plants had higher P concentrations and absorbed more P than non- mycorrhizal plants. Figure JO shows the relation between DM production and available P content ( Bray 11) of the soil after harvest, both for inoculated and noninoculated treatments. Curves were drawn visually through the points and arrows indica te "critica!'' P contents, defined as 95% of maximum yield. Although inoculated and non-inoculated plants reached the same maximum yield , it is clear that the presence of mycorrhiza markedly reduced the "critica!'' soil P-content. The critica! leve! of 15 ppm for mycorrhizal plants is only slightly above the leve! of 8-10 ppm obtained from field experiments, while the leve! of 190 ppm for non- mycorrhizal plants is unrealistically high. P applied (kg/ ha) • Noninoculated O Spore-inoculated .6. Root-inoculated Figure 9. E/Jet:t of mycorrhizal inoculation and severa/ levels oj applied P on dry maller production Jn tops oj cultivar M M ex 59. in a sterilized soil jrom C/A T- Quilichao. 67 Table 3. Ellect ol mycorrhsza l inoculation and P levels applied on mycorrhizal infection of cassava roots, P concentra tion of tops and total P u ptake by tops 1. P apphed Dcgree of intection l P concentration in tops (%) (~g . ha) NI ' Rl SI NI Rl o o 1.7 o 0.05 0.08 25 o 2.2 o O .o? 0.07 50 o 2.6 o 0.04 0.11 100 o 2.6 o 0.05 0. 12 200 o 2.4 o 0. 17 400 o 2.0 o 0 .06 0.17 800 o 1.5 o 0.09 0. 18 1600 o 1.0 o 0.15 0. 16 3200 o 1.0 o 0.20 0.25 • Culuvar M Mex 59 grown on stenlized soil from CIAT-Quilichao, in the greenhouse. ' Dcgree of infection . O • none, 3 • tugh number of hyphae and vesicles in roots. ' 1'1= oomnoculated. Rl=root-inoculated. S l.., spore-inoculated. 50 P in soil, Bray 11 (ppm) e lnoculated O Noninoculated P uptake by tops (mg/ plant) SI NI Rl SI 0.05 0.2 27.7 0.3 0 .05 0.5 25.5 0.6 0.08 0.3 41.3 1.5 0.3 49.4 0.09 81.0 2.0 0.09 1.2 73.0 2.8 0.06 3.4 86.0 3.4 0.08 58.3 75.3 27.5 0.2 1 90.5 118.3 93.5 Ftgure 10. Relation het,..,.een dry maller yields of three-monrh-old inoculated and noninoculated cassa\•a cultivar M M ex 59, and wd P conrent after harvest. (Arrows indicate critica/ P ie•·elsjor 95c;(; o/nwxm.um yield.) 68 Spore inoculation did nol result in an apparent root infeclion, nor in an improvement in plant growth, but inspeclion of the soil after harvest revealed a large spore populalion. Replanting cassava in these pots resulted in a marked improvement of growth at intermediate P levels. While infection from spore inoculation was much slower than from root inoculation, with time these methods will probably be equally effective. The results of this experiment corroborate the conclu- sion (CIA T Cassava Pro gr. 1979 Ano. Rep.) that without a mycorrhizal association cassava roots are extremely inefficient in P uptake, and indicate the great dependence of the crop on mycorrhiza when grown on low-P soils. Practica! implicalions of these findings are, however, not yel clear. Firsl, in a nonsterilized soil the native mycorrhizal population may be highly efficient, reducing the beneficia! effect of inoculation unless more efficient species and strains can be identified. Secondly, cassava is normally grown from stakes with a considerable reserve of nutrients (25-50 mg P 1 stake but less than 1 mg in rooted cuttings), which il can utilize lo grow an efficienl rool syslem before having lo rely on lhal syslem for P uptake. Initial responses to inoculation in cassava grown from stakes were small compared with those in cassava from rooted cuttings. lt remains to be seen what tbe long-term effect of inoculation is once tbe stake reserves are exhausted. Host specifity. To determine the degree of host specificity, mycorrhizal-infected cassava roots were used to inoculale seven specíes: maize, beans, cowpeas, rice, Andropogon gayanus, Srylosanthes guyanensis and cassava hybrid CM 91-3, all grown in sterilized soil from CIAT-Quilichao with O, 100, and 500 kg P/ ha applied. Table 4 shows the response lo inoculalion and P application of lhese species. All species except rice benefited markedly from inoculalion at the levels ofO and 100 kg P 1 ha, while cassava, beans, cowpeas and Sty losanrhes sp. also responded at 500 kg P/ ha. The lack of response in rice may indicate sorne host specificity, but is more likely due to lhe extremely fine and highly branched root system of this species. The lack of response to inoculation of Andropogon sp., and maize at 500 kg P 1 ha also reflects a well developed , highly branched and dense root system. As was observed in nutrient solutions (CIAT Cassava Progr. 1979 Ann. Rept.), without mycorrhizal infection cassava required much higher rates of P to attain normal growth than even such P-demanding species as beans. Using the ratio of DM obtained without inoculation over that obtained with inoculation as a measure of mycorrhizal dependence (Table 4), considering all P levels, cassava was the most mycorrhizal-dependent species, followed by Stylosanthes guyanensis. Without applied P, lhese two species again were most mycorrhizal dependent, although in reverse order; a t 100 kg P ¡ ha, Stylosanthes, A n- dropogon and cowpeas were even more mycorrhizal- dependent than cassava. On tbe other hand, with a mycorrhizal association, cassava was more P-tolerant (DM Po f DM p 5oJ than any of the species including rice. Thus, cassava's well-known ability to grow on low-P soils is dueto a large Preserve in its planting piece as well as to an effective mycorrhizal associatíon in many (but not necessarily all) low-P soils. Table 4. Effect of mycorrhizal inoculation and P application on dry matter in tops of seven crop species1• Species Dry matter in tops (g/ pot) Noninoculated lnoculated NJ P- 100 P-500 P-0 P-100 P-500 Cassava 0.34 0.72 0.54 4.33 14.21 16.36 Beans 1.1 1 3.44 8.29 3.08 18.79 25.01 Cowpea 0.96 0.64 13.65 2.60 20.68 36.32 Stylosanthes sp. 0.08 0.08 2.74 1.25 9.33 12.20 Andropogon sp. 0. 15 0.39 34.24 1.26 16.67 32.18 Maize 1.19 8.74 59.35 4.84 34.75 53.57 Rice 3.79 26.63 30.60 3.83 22.36 31.23 1 Crops grown in sterilized soil from C IAT-Quilichao, in the greenhouse. Fertilized with either 100 or 500 kg P/ ha. ' Mycorrhi7.al depcndcnce is in terms of the ratio: Dry man er produced without inoculation Ory mattcr produced wíth inoculation Mycorrhizal dcpendcnce ra tio l P-0 P-100 P-500 Mean 0.08 0.05 0.03 0.05 0.36 0.18 0.33 0.29 0.37 0.03 0.38 0.26 0.06 0.01 0.22 0.10 0. 12 0.02 1.06 0.40 0.25 0.25 1.11 0.54 0 .99 1.19 0 .98 1.05 69 Errata Page Columm Element Pnntcd : Should be: 6 Figure 2 M (ol S9 M Mex 59 6 2 Figure 3 M Col 59 M Mex 59 6 2 figure 3 1 \ D ( I'