Potato yield and quality as a function of the plant density

The effects of potato plant density on yield quantity and quality were investigated at the Hankkija Plant Breeding Institute from 1971 73, using seed rates of 1600, 3200 and 4800 kg/ha, and seed sizes of 40, 80 and 120 g. The varieties used were Ijsselster and Record. The number of stems per m 2 rose with increasing seed rate and with increasing seed size. Stem number increased with seed rate faster for small seed than for large. The response in stem number was greater for Ijsselster than for Record. The number of stems per seed tuber fell as the plant density rose. The number of tubers per m 2 altered in the same direction as the number of stems, but less responsively. The reason for this was that the number of tubers per stem decreased with increasing plant density. The tuber yield showed a continual increase with increasing plant density. At the lowest stem densities (less than 20 —25 stems/m2 ) small seed gave better results than other sizes, but at the higher plant densities, the importance of seed size faded away and the yield was dependant on the plant density alone. Net yield (gross yield 2 x seed rate), however, was higher the smaller the seed used, whatever the stem density. Tuber size decreased when plant density increased, the proportion of large tubers diminishing most, especially when small seed was used. The proportion of small tubers altered more for Ijsselster than for Record. Seed size and seed rate did not on average significantly affect the proportion of Class I potatoes, though small seed gave results slightly better than other sizes. The starch content of the yield rose when the seed rate was increased (16.0 16.3 16.5 %) and fell with increasing seed size (16.5 16.2 16.1 %). The maximum variation was 15.8 16.7 %. The specific gravity distribution improved with increasing plant density. Raw discolouration of the tubers did not alter significantly as the plant density rose. Blackening of the tubers decreased with increasing plant density when small and medium sized seed were used. Mealiness of the tubers was somewhat higher for small seed than for other seed sizes. Potato plant density is nowadays taken to mean the number of main stems per unit area (Bleasdale 1965). In the past, plant density has been regulated by choice of seed size and spacing ignoring factors relating to the production of stems on the tubers. However, the results thus attained are rather rough, since the capacity of a given seed size to produce stems varies considerably with variety, seed quality and treatment (Dent and Hal-

Potato plant density is nowadays taken to mean the number of main stems per unit area (Bleasdale 1965).In the past, plant density has been regulated by choice of seed size and spacing ignoring factors relating to the production of stems on the tubers.However, the results thus attained are rather rough, since the capacity of a given seed size to produce stems varies considerably with variety, seed quality and treatment (Dent and Hal- kon 1969).Likewise weight of the tubers and number of stems produced by them are not linearly related (Varis 1973 a).
For this reason the definition given above has been adopted, and plant density has been assessed from the capacity of seed tubers to produce stems.
However, this requires a better understanding of the factors affecting the formation of stems from seed potatoes, and of the methods used for deter- mining stem numbers, than we have at present (Hagman 1973).
It has been shown that under intensive cultivation a moderate degree of change in plant density has hardly any effect on gross yield, and even less on net yield, since the potato has a fairly good capacity for using spare room either by increasing the number of tubers per stem, but mainly by increasing tuber size (Dambroth and Pätzold 1969).There is less infor- mation as to how variations in plant density are reflected in quality of the crop, excepting variation in tuber size (Warren 1958, Bleasdale and  Thompson 1969).
An attempt is made in this study to investigate the effects of seed size and rates upon plant density, and the effects of plant density upon yield and quality.

Materials and methods
The material comprises four experiments from 1971-73.Three of the experiments were made at the Anttila Experimental Farm, and one was made at the field site in Viskaali, Muhos in 1972

Hekord
Ijsselster 4800 » 120 * The treatments 1600-40, 3200-80 and 4800 -120 contained the same number of tubers per unit area.There were two replicates, the plot size was 10.05 m 2 and the experi- mental layout a factorial design.The soil type at Anttila was loam under- laid by heavy clay, at Muhos finer finesand.The fertilizer rate was 1200 kg of a chloride-free complete N-P-K fertilizer (7 -ll 12) per hectare, drilled across the plots.The ridges were spaced at 67 cm and within the ridges the following planting distances were used:   The average date of planting the experiments was the 26th of May, that of lifting the 14th of September.Normal cultivation techniques were used.
A »Nosto-Juko» potato harvester was used to lift the crop, and the remain- ing small potatoes were gathered by hand.
The following determinations were made upon the stands and harvested tubers; 1. Number of main stems/m 2 .The stems were counted at the end of August.2. Number of tubers/m 2 .The tubers were counted when the yield was weighed.
4. Tuber size distribution (%).The yield was sorted using 40 and 55 mm sieves.5. Class I yield (%).The proportion of the yield meeting Class I standards as specified in the Finnish statute for table potatoes was assessed after sampling for starch con- tent determinations.
7. Starch content.For starch determinations, 5 kg samples were taken from the plot yield.The weight in water was converted to starch-% according to the tables Of Hals & Bucholz.
The analyses of variance used for the statistical treatment of the results were calculated in the ADP department of the Hankkija Cooperaticve Centre.Significance levels for the results are shown in the usual manner.

Results and discussion
The results of the analyses of variance are shown in Table 1.

Number of stems
The mean number of stems varied from trial to trial (23 21-23 -32   stems/m 2 ).The value for 1973 trial clearly deviated from values for the other trials.There were also differences between the varietal means (Record 21, Ijsselster 32 stems/m 2 ), varying yearly (AD).An increase in seed size and seed rate increased the number of stems in every trial (B, C, AB and AC).
The interaction between seed size and rate was significant (BC) as were the interactions seed size X variety (BD), seed rate X variety (CD) and seed size X seed rate X variety (BCD, Fig. 1).There was yearly variation in these interactions (ABD, ACD) except for the interaction BC, which was similar in all the experiments.
The number of stems increased with seed rate more rapidly for small seed than for large.The effect was readily discernible with Ijsselster.The Table 1.Analysis of variance of the effects of plant density on yield and quality of potatoes.stem number was less responsive to seed rate in the case of Record and the rate of increase of stem number was less dependent upon seed size.In nearly all cases the number of stems increased almost linearly.It is apparent from Fig. 1 that a good plant density is more readily achieved by reducing seed size and spacing than by increasing seed size.
The reason for this is that small or rather small tubers have, weight for weight, more shoots than large ones (cf.Varis 1973 a).Naturally, if seed tuber spacing is held constant, a higher plant density is achieved with large than with small seed tubers.
An increase in seed rate, however, caused a reduction in the number of stems per seed tuber, apparently because of competition between adjacent hills (Fig. 2).The stem number fell with increasing seed rate, particularly when large seed was used (cf.Morris 1967).

Number of tubers
The number of tubers varied in the four trials (52 51-46-68 tubers/m 2 ).
An increase in seed rate and in seed size increased the tuber density (B and C).
Numbers of tubers were different for the two varieties (Ijsselster 68, Rekord 40 tubers/m 2 ).The interactions for number of tubers were fewer than for number of stems.Thus the interaction seed size x seed rate (BC) was not significant, nor were any of the second degree interactions.
Numbers of tubers for both varieties were greatest when small seed was used (Fig. 3).Tuber numbers increased when the seed rate was raised, though not in proportion to stem numbers (cf.Table 2).
When small or medium-sized seed was used, the number of tubers per stem was on average a bit larger than when large seed was used (2.2 2.2 -2.1).The number of tubers per stem fell when the seed rate was increased (2.5-2.2-1.8)(cf.Houghland and Akeley 1959, Dambroth and  Pätzold 1969, Varis 1973 a).

Yield of tubers
The yield of tubers varied in a normal manner from trial to trial (A).The mean yields of the varieties were different (Ijsselster 26.96 tons/ha, Record 25.35 tons/ha).The seed rate influenced yield (20.36-27.3230.80 tons/ha) as did seed size (28.88 25.60-24.00tons/ha).Only the interac- tion trial X variety (AD) was significant.There was no significant interac- tion of seed rate and seed size upon yield.
Figure 3.The effects of seed size and rate on the tuber number per m 2 Small seed produced higher yields of both varieties (Fig. 4).The yield increased as the seed rate was increased, though not linearly.The same trend was shown by Pohjonen and Paatela (1974).
The yield data have been plotted as a function of stem densities in Fig. 5.When the stem density increased from 20 to 40 stems/m 2 , the yield rose almost linearly by about 7 tons, or 350 g for each additional stem.
At low stem densities, small seed gave better yields than larger sizes.However, at higher stem densities the effect of seed size lost much of its importance, and the yield seemed mainly dependent upon the number of Figure 4.The effects of seed size and rate on the tuber yield.
Figure 5.The effect of stem density on the tuber yield and quality, stems per unit area (cf.Holmes 1966).The varieties differed in that Record produced a higher yield per stem than Ijsselster, although Record gave on average a poorer total yield due to its smaller number of stems.
As well as gross yields, net yields are shown in Fig. 6.Net yields are estimated using the formula gross yield 2 X seed rate = net yield.The smallest seed gave the best result at all stem densities.The reason for this is that small seed has the most sprouts and tubers for a given weight (cf.Varis 1973 a).

Size distribution
All the treatments had a significant effect upon size distribution.There was yearly variation in the effects of seed rate and seed size (AC and AD).Also the interaction seed rate x variety (CD) was significant.
The proportion of large tubers in the yield fell both with increasing seed rate and with decreasing seed size.(Fig. 7).This is the predictable consequence of changes in plant density and yield (cf. Dambrot and Pätzold   1969).According to Thompson and Taylor (1974), as plant density in- creases, tuber size soon levels off at a density of 70 stems/m 2 .The results of this study show the same trend.
An increase in the seed rate increased the number of small tubers for Ijsselster more than for Record (Ijsselster 11 16-18 %, Record 4-5 5 %).
The two factors determining potato yielding capacity are the number of tubers and their size.To produce tubers of a given size, one must adjust the plant density on the basis of the expected yield (cf. Varis 1971).In the regulation of plant density, not only seed size but also the variety Figure 6.The effect of stem density on the net yield (net yield = total yield 2 X seed rate).
must be considered, as varieties differ in their capacity to form stems and tubers (Varis 1973 b).The simplest way of determining the planting rate would be to use the number of sprouts capable of developing shoots per tuber planted (cf.germination percentage of cereals), always provided that a reliable method of determination could be devised (Hagman 1973).

Proportion of Class I yield
The proportion of Class I yield varied greatly from trial to trial (54 83 %).There was also a difference between varieties (Ijsselster 78 %, Record 62 %).Seed rate and seed size, however, caused on average no significant differences.In fact, the effects of seed rates were conflicting in different trials (AD).The interaction seed rate X seed size was slightly significant (Fig. 5).Small seed produced the largest yield of class I potatoes, but the increase in class I yield with increasing plant density levelled off faster than did the gross yield, and usually reached a plateau at a stem density of 40 -5O stems/m 2 , depending upon seed size.
The starch content rose with increasing seed rate (16.0-16.3-16.5)and fell with increasing seed size (16.5-16.2-16.1 %).Judging from these trends, it would seem that the high plant density attainable by the use of small seed is favourable for starch production (cf.Bleasdale and Thompson  1969).

Specific gravity distribution
The specific gravity distribution varied yearly and with variety.An increase in seed size caused a wider distribution, and a higher seed rate narrowed the distribution (Fig. 9).In general, the higher the mean starch content of the yield, the larger the proportion of tubers approaching the genetic starch content maximum for a variety, and the larger the proportion of tubers falling into high specific gravity classes.The specific gravity distribution of Record was less favourable as regards yield quality, than that of Ijsselster (Fig. 10).

Raw discolouration
There were yearly and varietal differences in raw discolouration.Neither seed size nor seed rate caused on average any significant changes, though the results in different trials were conflicting (AB, AC).The interaction BC was slightly significant (Fig. 11).Large seed behaved differently from other sized seed, which observation is difficult to account for.

Blackening
Blackening also varied yearly and with variety (A, D).Also an increase in seed rate reduced blackening slightly (C), when small and medium sized seed was used (Fig. 11).Bleasdale and Thompson (1969) arrived at the same result.

Sloughing
The mealiness of potatoes was estimated from the extent of disintegra- tion on cooking.Only yearly variation (A) was found.A slight interaction BC indicated that small seed produced more mealy potatoes than larger seed sizes, especially when intermediate seed rates were used (Fig. 11).In studies of Bleasdale and Thompson (1969), an increase in plant density aggravated the tendency of potatoes to slough.

Conclusions
The results on potato plant density presented here, as well as those els- where, show that the optimum plant density is achieved most easily by using fairly small seed.The actual optimum plant density to be used depends upon the cultivation conditions and upon variety.However, under Finnish conditions it would seem relatively high, 30-4O main stems per m 2, com- pared with the recommendation of 20-25 main stems per m 2 used in Hol- land, for instance (van der Zaag 1973).This, from our point of view un- favourable situation may be due to the fact that the ability of our seed to produce tubers per stem is poorer than that of seed in Central Europe.
On the other hand, the shortness of our growing season involves a need of a large assimilating leaf area relative to the size of the developing yield, so that the quality of the yield can attain a high standard.
From what has been stated, it follows that under Finnish conditions, the seed requirement per unit planting area is larger than in Central Europe.
To mitigate against this disadvantage, we should continue to investigate all the factors associated with the characters of seed potatoes, and of growing conditions, so that large seed rates and high costs could be reduced without any sacrifice in yield, and particularly in quality.

2Figure 1 .
Figure 1.The effects of seed size and rate on the stem number per m 2 .

Figure 2 .
Figure 2. The effects of seed size and rate on the stem number per seed tuber.

Figure 7 .
Figure 7.The effect of stem density on the size distribution.

Figure 8 .
Figure 8.The effect of stem density on the starch yield and starch content.

Figure 9 .
Figure 9.The effects of seed size and rate on the SG distribution

Figure 10 .
Figure 10.The SG distributions of the cultivars Ijsselster and Record, . The treatments were as follows:

Table 2 .
The effect of the seed rate and size on the tuber number per stem.