Assessing the invasive status and potential for biological control of Bilberry , Myrtillocactus geometrizans.

Ekhona Zozo1*, Iain Paterson1, Kanyisa Jama2 1Centre for Biological Control, Department of Zoology and Entomology, Rhodes University, Grahamstown, South Africa 2South African National Biodiversity Institute (SANBI), East London, South Africa *[email protected]

ABSTRACT Alien invasive are generally a problem in the areas they invade around the world. This is because of the negative impacts they usually have in the ecosystems they invade. Amongst the most problematic alien invasive plants are those in the Cactaceae family. These are found in South Africa where they have a negative impact on indigenous fauna and flora. Biological control of cacti in South Africa has provided solutions to the problem of these alien invasive cacti. One of the currently emerging alien invasive cacti is Myrtillocactus geometrizans (Mart. ex Pfeiff.) Console. This cactus was probably introduced into South Africa through horticultural trade. This cactus is likely to become another problematic one because it is closely related to some of the most damaging cactus species. Hypogeococcus sp. (Hemiptera: Pseudococcidae) has been used in Australia and South Africa to control alien invasive cacti such as Harissia martinii and Cereus jamacaru. It is reported as having done a good job in clearing vast areas of those cacti. Here we investigated the invasive status of this emerging cactus and determined if the , Hypogeococcus sp. would be a damaging biological control agent to M. geometrizans. Determining the invasive status of the infestations was done by looking at the population dynamics of each infestation. To do this we found the biggest/oldest in each infestation, measured parameters of this plant, i.e. the height, canopy diameter, the number of branches, and whether each plant had reached an age that was reproductively viable by checking if there were any flowers, fruits or flower buds. To determine if Hypogeococcus sp. would be a damaging control agent, M. geometrizans, and four other cactus species, namely, Cereus jamacaru DC., balansae (K.Schum.) N.P.Taylor & Zappi, Harrisia martinii (Labour.) Britton and Hylocereus undata (Haw.) Britton & Rose, was planted plant pots. These were all replicated ten times. On each plant, parameters which were measured are: number of branches, length of each branch, height of the whole plant, the circumference of the trunk base, the circumference of the thickest branch before the top, and the diameter of the canopy of the whole plant. A photograph of the plant, next to a ruler was taken. The biological control agent, Hypogeoccocus sp., was introduced by making a small cut on the thickest branch of each plant. Then, one Hypogeococcus sp. gall, on each plant, on a sosatie stick was introduced on the branch on the area cut and left there. The results we obtained show that the infestations surveyed in this study are small since they are easy to quantify, these populations are, however, spreading and proliferating at a potentially high rate. This suggests that, similarly, the other populations detected around the country may also be growing and spreading. Myrtillocactus geometrizans is a high-risk species which should be kept under Category 1a of the NEMBA legislation. This species also has a high percentage of viable seeds and is at an advantage since it can spread through vegetative plant parts and could also potentially germinate from dispersed seeds from fruit eaten by birds, other animals and human beings. We were, unfortunately, not able to obtain and quantify any significant results from the biological control agent damage test. We do, however, still view Hypogeococcus sp. as a good potential candidate for the control of M. geometrizans—we suggest further studies on this both at the individual and population levels. There is still a need to come up with an eradication or control strategy for this alien invasive species before it spreads beyond control and becomes a problem.

Keywords: Emerging, mealybug, weeds, alien invasive, population dynamics.

INTRODUCTION Invasive cacti in South Africa Various cacti species have been documented as invasive in South Africa; these include many species from the Opuntia (Paterson et al. 2011). These species came into the country in different ways, some were introduced unintentionally, whilst others were introduced intentionally through horticultural trade and for commercialization of, for example, their fruit (Novoa et al. 2015b; Novoa et al. 2017). Many of these alien cacti have subsequently become very problematic in South Africa because they cause negative change in the areas they invade, a process referred to as biological invasion, which poses a threat to the conservation of natural resources and biodiversity (Simberloff et al. 2013).

Cactaceae are considered as amongst some of the most damaging invasive alien plants in South Africa (Kaplan et al. 2017). Some of their negative impacts include reducing the carrying capacity of land which could be used for grazing for livestock and wildlife; they also reduce indigenous plant diversity (van Wilgen et al. 2004). Animals which encounter cactus infestations are usually harmed and/or killed—leading to the compromising of the quality of natural resources such as wool and a reduction in native animal biodiversity (van Wilgen et al. 2004).

Though it is true that some alien invasive cacti are found in very limited populations, they are likely to become problematic in the future because they have similar characteristics to those cacti which are currently problematic.

Biological control is a method of control which has been used for alien invasive cactus control in South Africa. The first intentional biological control program was done in 1913 when Dactylopius ceylonicus (Green) (Hemiptera: Dactylopiidae) was used to control Opuntia monacantha Haw. (Cactaceae) (Paterson et al. 2011). Since then, the work of biological control has continued in the country on various cacti using different agents. Many biological control programs have been very successful (Paterson et al. 2011; Moran et al. 2013). In some cases, there are biological control agents which control multiple species, e.g. Hypogeococcus sp. controls Cereus jamacaru, Harrisia balansae and H. martini (Paterson et al. 2011; Moran & Zimmermann 1991).

There is a need to assess cactus species with limited distributions, so that it can be determined if they are likely to become problematic and whether their eradication or control is possible. It is a requirement of the National Environmental Management Biodiversity (NEMBA) Act 10 of 2004 that a control strategy be developed for each alien cactus and to determine whether eradication is feasible, or chemical/mechanical or biological control should be used instead.

Myrtillocactus geometrizans The Bilberry Cactus, Myrtillocactus geometrizans (Mart. ex Pfeiff.) Console, is a cactus species that obtains a height up to 4 or 5 metres (Walters et al. 2011). It has a short trunk with numerous, branches that curve upwards, these branches have a blue-green appearance and are 6 to 10 cm in diameter (Walters et al. 2011). Branches have an average of 6 ribs and areoles are separated by between 5 to 30 cm (Walters et al. 2011). Each areole has one central, dagger-like spine, measuring 1 to 7 cm that is black in colour (Walters et al. 2011). The radial spines vary in number from five to nine and are 2 to 10 mm long (Walters et al. 2011). During the flowering season, the tree produces creamy or greenish white flowers, that develop into dark purple or red, globose fruit which are about 2 cm in diameter (Walters et al. 2011).

This cactus species is native in Mexico, a country characterised by a semi-arid to arid climate (Hernandez-Lopez et al. 2008; Novoa et al. 2017). Like many cactus species which have been introduced into countries around the world, this species is likely to have been introduced into South Africa through the horticultural trade as an ornamental plant (Novoa et al. 2015b). Myrtillocactus geometrizans is considered as an emerging weed in the country. Emerging weeds are alien invasive plants that are in the early stages of their invasion (Olckers 2004). The species is currently listed, in Category 1a, for control under the National Environment Management: Biodiversity Act, Alien and Invasive Species Regulation (NEM:BA) (Novoa et al. 2015a).

Nine areas in South Africa have been reported as invaded by M. geometrizans and some of these areas are very densely infested. The cactus can spread within a site and to new sites from, seeds dispersed after the fruit has been eaten by animals; as well as from detached stems (Chuk 2010). It is likely that M. geometrizans will become problematic if it is not controlled. This assumption is based on other well-studied alien invasive cacti which have been shown to have negative impacts on native fauna biodiversity by, for example, reducing the carrying capacity of land which could be used for grazing of livestock and wildlife (van Wilgen et al. 2004). It is therefore important to either eradicate or implement the control of M. geometrizans before it becomes a problematic alien species in the South Africa.

The other cacti Cereus jamacaru Cereus jamacaru DC is native to Brazil (Braun et al. 2017). It grows on sandy to stony soil and on rocks in various ecoregions such as caatinga and agreste which are desert ecoregions (Braun et al. 2017). Its major threat in the native region is habitat loss—its habitat is being converted for use in agriculture and cattle ranching (Braun et al. 2017).The species is used in its native country to feed cattle, its fruit is edible, it is used as a hedge and was, in the past, used as timber for construction (Braun et al. 2017).

Cereus jamacaru is an alien invasive in South Africa. It has been recorded in all provinces of the country with large infestations occurring only in Gauteng and Limpopo (Klein 1999). Hypogeococcus pungens Granara de Willink (Hemiptera: Pseudococcidae), and the stem borer Nealcidion cereicola (Fisher) (Coleoptera: Cerambycidae) were released for the biological control of this cactus and they established well on it (Paterson et al. 2011). Both biological control agents have been very damaging to the cactus and thus have been effective in keeping it under control (Paterson et al. 2011).

Harrisia balansae and H. martinii Harrisia cactus species are perennial plants with fleshy, spiny stems (The State of Queensland 2017). These stems are jointed and make tangled mats which are about a metre in height (The State of Queensland 2017). These weeds originated from Argentina and Paraguay (Tomley & McFayden 1984; Klein 1999; Klein 2002a).

Harrisia martinii was, at one point, considered as a weed of restricted distribution and minor importance in South Africa (Moran & Zimmermann 1991) but this cactus became more significantly spread around the country (Klein 1999). This shows how the population dynamics of this important alien invasive family can change within a short period.

Two biological control agents have been used to control H. martinii in South Africa, these are the cerambycid Alcidion cereicola (Fisher) (Coleoptera: Cerambycidae) and the pseudococcid Hypogeococcus festerianus Granara de Willink, 1981 (Klein 1999; Klein 2002). The mealybug was initially released in 1983 (Moran & Zimmermann 1991).

Hylocereus undata Commonly known as the dragon fruit cactus, Hylocereus undata (Haw.) Britton & Rose, believed to be native to the Pacific side of Guatemala, El Salvador, Costa Rica and southern Mexico but there has been no confirmation of this (Durán et al. 2017). It has long been cultivated in many countries around the world and is the most cultivated species in this genus (Durán et al. 2017). The species often escapes areas of cultivation and becomes naturalized in the areas it proliferates to (Durán et al. 2017).

Currently the species is widely distributed around the world (Durán et al. 2017). It can be found in Central America, tropical South America, the tropical forests of Mexico, various Caribbean islands, and southeast Asia (Durán et al. 2017).

Hylocereus undata is a shade-tolerant lithophyte or hemi-epiphyte which can be found in the low tropical deciduous forests, riparian vegetation, thorn forests, tropical deciduous forests, and thorn scrubs, of the areas in which it becomes naturalized (Durán et al. 2017). It is used in various ways, including as a medicinal plant (Biswas et al. 2009; Perez et al. 2005), its fruit is also edible to human beings and has been commercialized (Durán et al. 2017).

The potential biological control agent The Hypogeococcus sp. belonging to the Pseudococcidae and commonly known as the Harissia cactus mealybug has many times in the past, been identified as H. festerianus and H. pungens (Klein 2002b; Segarra-Carmona et al. 2010). But there has been debate about the species to which this mealybug belongs (Klein 2002b).

This pseudococcid is a polyphagous feeder on columnar cacti around the world (Segarra-Carmona et al. 2010). Its impact on the cacti is very significant such that it is a dangerous pest in countries where it is invasive and is not necessary as a biological control agent (Segarra-Carmona et al. 2010). In the Central and North America and Caribbean countries such as Puerto Rico, the mealybug is a threat to native cacti biota (Carrera-Martínez et al. 2015).

In Australia and South Africa, however, this mealybug has been used to control alien invasive cacti, and reports state that it has done a good job in clearing vast areas of those cacti (Segarra-Carmona et al. 2010; Tomley & McFayden 1984; Moran & Zimmermann 1991; McFadyen and Tomley 1981a). The mealybug has been used to control cactus species such as H. martinii and C. jamacaru (Segarra- Carmona et al. 2010; Paterson et al. 2011). In C. jamacaru it reduces the production of fruit and leads to both seedling and large plants’ death (Paterson et al. 2011). These characteristics of this mealybug led us to selecting it as a potential biological control agent for M. geometrizans.

Population dynamics Sutton et al. (2018) in their evaluation of the effectiveness of Hypogeococcus sp. as a biological control agent of C. jamacaru, used age frequency distributions to visualize the population dynamics of the alien invasive cactus populations. In these distributions, age classes of a population are plotted against the number of individuals in each age class (Sutton et al. 2018; Paynter 2005). Three scenarios exist for population dynamics, as shown by age frequency distributions. The first one occurs under optimal conditions, here, the plants have a constant spreading rate and a mortality rate which is either constant or decreasing with age (Agren & Zackrisson 1990; Sutton et al. 2018). The age frequency distribution of these populations is reverse-J shaped and thus shows a population which is able to self-regenerate and is expanding (Sutton et al. 2018).

Sutton et al. 2018 also displays age frequency distributions which are J-shaped. These are decreasing populations where there is a low spreading rate and a high mortality rate of seedlings (Sutton et al. 2018). Lastly, the rate of spread and mortality can be relatively equal. In such populations the age frequency distribution is normally shaped. Here, the populations are stable (Sutton et al. 2018). Thus, using age frequency distributions, we are able to determine whether a particular species is spreading or not. We can also use these distributions as evidence for whether a population is becoming invasive or not.

Methods of Control Alien invasive cacti currently with limited distribution, such as Cylindropuntia leptocaulis (DC.) F.M. Knuth and M. geometrizans need to be assessed for their invasiveness. It is important to know the dynamics of these emerging invasive species populations and thus determine which method(s) of control is/are suitable for them and whether eradication is feasible or not. Such studies are in line with the NEMBA National Strategy for dealing with biological invasions in South Africa recommendations (Zachariades et al. 2017). These invasive species are labelled as ‘proactive targets’ and should be of interested focus for control lest they, like the cacti species preceding them, become problematic and introduce costs which could have been avoided.

Methods of control available for these proactive targets are chemical, mechanical or biological control. These methods could be done in isolation from one another or applied in an integrated manner. Examples of invasive cacti which have been controlled are using an solely-biocontrol or integrated approach and which are under complete and substantial control include C. leptocaulis, Opuntia humifusa (Raf.) Raf., Harrisia balansae (K.Schum.) N.P.Taylor & Zappi and H. pomanensis (F.A.C. Weber in K.Schum.) Britton & Rose (Zachariades et al. 2017).

Aim and objectives The aim of this study was to determine the invasive status of M. geometrizans and the potential to use Hypogeococcus sp. as a biological control agent of this cactus.

Objectives were to determine (a) the total population size of the three selected M. geometrizans infestations, (b) the population dynamics of this species. Is it increasing, stable, or decreasing? (c) the risk of having this species in South Africa—including by conducting a Tetrazolium seed viability test— and (d) whether Hypogeococcus sp., is a damaging biological control agent against M. geometrizans.

MATERIALS AND METHODS Study site The Southern African Plant Invaders Atlas (SAPIA) 2018 database records were used to determine the known distributions of M. geometrizans in South Africa. From the nine known populations in this database three were selected for this study; the first one in Graaf-Reinet, the second at Damsedrif Farm in Baviaanskloof and the third at Prince Albert (Fig. 1).

Population size and dynamics To determine the population size of each infestation, the total number of individual trees was determined by counting them. Then, for the population dynamics, the largest tree was first selected in each population—this tree was assumed to be the earliest and oldest colonizer of the infestation. From this oldest tree, which was recorded as plant 1, all distances to other plants in the infestation were measured using a tape measure. Other plant parameters which were measured and recorded on the data sheet (Appendix 1) were plant height, the tree canopy diameter, and the number of branches or stems on each tree. We also determined whether each plant had reached an age/size that was reproductively viable by checking if there were any flowers, fruits or flower buds. Risk Assessment Due to there being no knowledge of an existing risk assessment for M. geometrizans, a risk assessment, using the Weed Risk Assessment Model developed by Pheloung et al. 1999, was done. This model was modified to fit the South African context (Appendix 2a) by modifying a few questions. This Weed Risk Assessment has 49 questions which are based on the main characteristics and impacts of invasive species (Pheloung et al. 1999). The questions fall into three sections namely, History, Biogeography, Biology and Ecology of the plant (Pheloung et al. 1999; Riddin et al. 2016).

The Weed Risk Assessment scoring sheet (Appendix 2b) provides a score for each response to the questions (Pheloung et al. 1999; Riddin et al. 2016). When the scores have been combined, the total score determines one of three outcomes: a score less than 1 means the alien plant is not invasive and can be allowed into the country, one between 1 and 6 means the weed needs to be evaluated further, then a score greater than 6 means the weed is invasive (Pheloung et al. 1999; Riddin et al. 2016). This model for risk assessment has been shown to be accurate across a broad geographic scale (Gordon et al. 2008; Riddin et al. 2016).

The Tetrazolium seed viability test The Tetrazolium seed viability test or assay is a quick evaluation of seed viability and germinability (Verma & Majee 2013; Wharton 1955; Porter et al. 1947). This assay takes advantage of the fact that all tissues respiring can convert the colourless compound—2,3,5 triphenyl tetrazolium chloride—to a carmine red/pink coloured water-insoluble formazan (Verma & Majee 2013). Thus, living cells are stained red or pink at the end of the assay (Verma & Majee 2013).

We collected fruits from M. geometrizans at the Baviaanskloof infestation. Five of these fruits were used for the Tetrazolium seed viability test. In this test we removed the seeds from each fruit. Seeds from each fruit were counted, the number recorded and placed in five separate glass vials; each vial was labelled. The vial with fruit one was labelled “F1” and the one with fruit five was labelled “F5”. The others were also labelled according to the fruit number whose seeds they contained. Each of these vials was filled with water to soak the seeds overnight. The following day, the seeds were all cut in half, on half was thrown away and the remaining half was used for the test.

The water from the vials was removed and halves used for the test were placed back in their respective vials and 2,3,5 triphenyl tetrazolium chloride solution (tetrazolium chloride solution) was added soaking the seeds. Each glass vial was covered externally with aluminium foil to avoid the entrance of light which would tamper with the results. The vials were then placed in an oven set at 50°C for 15 hours. After this time, the liquid was drained, the half-seeds removed. Then the number of seeds showing a positive result—pink/red staining were counted, recorded and a total percentage of viable seeds was calculated. Biological Control Four cactus species seedlings, namely, C. jamacaru, H. balansae, H. martinii and Hylocereus undata together with M. geometrizans were collected from the field and planted in pot plants. These cactus species were selected because they are currently being controlled by Hypogeococcus sp. (Paterson et al. 2011; Moran & Zimmermann 1991). The damage which this biological control agent inflicts on them is varying: C. jamacaru is damaged to a high degree, H. balansae and H. martini are also significantly damaged but to a lesser degree and H.undata receives comparably lower damage from Hypogeococcus sp. Therefore, we endeavoured to compare the damage which would be inflicted on M. geometrizans and compare this to the other cacti species damage to assess whether M. geometrizans could be controlled by Hypogeococcus sp.

The planted cacti were weeded and watered as necessary and allowed to grow for a few weeks. Before introducing the biological control agent, these parameters were measured and recorded from each plant: the number of branches, length of each branch, the height of the whole plant, the circumference of the trunk base, the circumference of the thickest branch before the top, and the diameter of the canopy of the whole plant. A photograph of the plant, next to a ruler was taken. Then, the biological control agent, Hypogeoccocus sp., was introduced by making a small cut on the thickest branch of each plant. Then, one Hypogeococcus sp. gall, on each plant, on a sosatie stick was introduced on the branch on the area cut and left there.

Statistical Analyses Statistical analyses were done using the Statistica.Ink software. The Shapiro-Wilks and Kolmogorov– Smirnov tests were used to test the data for normality. The Kruskal-Wallis H-test was used for our non-parametric data. A simple correlation between the size of each individual and distance from the first plant, was done for each site. The Product-Moment and Partial correlation methods gave us the p-values which indicated whether the correlation was significant or not. Note: plant size is the product of the two measured diameters. The number of branches in each site was categorized into groups and a one-way ANOVA was done together with a Kruskal-Wallis H-test. An analysis was also done to determine the size at which the first reproductively active fruits occurred.

RESULTS Population size From our three study sites, Damsedrif farm in Baviaanskloof had the highest population of M. geometrizans with 71 individual trees, it was followed by the Graff-Reinet site where there were 56 trees and lastly the Prince Albert site which had 54 trees. The average size of trees at the Baviaanskloof site was 108.51m2, at the Graaff-Reinet site it was 185.33m2, while at the Prince Albert site it was 243.86m2. The number of adults for Baviaanskloof, Graaff-Reinet and Prince Albert was 14, 23 and 40, respectively. The total number of trees from these sites was 181. Note: Refer to Fig, 2 and Table 1.

Population dynamics The correlation plots (Fig. 3-5), showing correlation between plant size and distance from each site’s largest tree, gave a negative correlation for all three sites. The Product-Moment and Partial correlation methods reported insignificant (>0.05) p-values for the Baviaanskloof and Prince Albert sites. But the p-value was significant (<0.05) for the Prince Albert site (p-value = 0.011). This may be due to the fact that many of the Prince Albert data points fit the trend line compared to the other sites. Note: tree size is the product of both the diameters we measured whilst collecting the raw data.

The size frequency histograms showing the distribution of the categorized numbers of branches in each site (Fig. 6) show that the highest number of plants in each infestation are the ones with a relatively small numbers of branches—the ones between 0 and 50 branches. These reverse-J shaped histograms show us that these populations are spreading and a One-way ANOVA by ranks and Kruskal-Wallis test showed that these results are significant ((2, N= 181) =27.18674 p = 0.00).

Table 2 shows the smallest sizes at which the trees from each site started to be reproductive, the number of branches each of those trees had and how many individual trees were reproductive in each site at the time the data was collected. Comparing the three sites we see that the smallest tree size reproductive was 104m2 this was observed at the Graff-Reinet site. This tree had only 26 branches. This site was followed by the Prince Albert site with 28 reproductive trees, the smallest one being 118m2 with only 34 branches.

Risk Assessment The total risk assessment score was 29 (Table 3). This high score was due to the fact that this cactus is a naturalized environmental weed which produces spines and thorns, it can grow in dry soils, produces viable seeds, can propagate itself by vegetative parts, its propagules have been dispersed both intentionally and unintentionally by birds, other animals and human beings. It also produces a large number of seeds in a single fruit and has many fruits in one stem. The Agricultural and Environmental scores were also high: 24, 21, respectively.

The Tetrazolium seed viability test The total number of seeds we obtained from the collected M. geometrizans fruits was 388. Out of these, 266 (68.56%) (Table 4) seeds tested positive for the Tetrazolium seed viability test.

Biological Control test After being grown for some weeks, we recorded quite high parameter measurements for the cacti seedlings (Table 5). Despite this, however, we did not get any quantifiable results for the biological control test because the time between the introduction of M. geometrizans galls and the end of the experiment was insufficient, and the seedlings had not grown to a substantial degree.

DISCUSSION Though the findings of this study indicate that the surveyed M. geometrizans infestations are not very densely infested, the individual trees in each population can be found at quite sparse distances from one another. This is an indicator of spread over the landscape which could mean a decrease in land carrying capacity for agricultural activities such as livestock grazing, it could also lead to native flora being out-competed for resources such as land space, nutrients, sunlight and water (Goeden et al. 1967).

The M. geometrizans infestations are spreading. This finding is supported by the negative correlation between plant size and the distance from the oldest/largest trees of the infestations. This means that as one moves from the biggest tree—which was more-or-less at the centre of the population—the size of plants decreased. Big plants can be found at and/or near the centre, while towards the periphery seedlings can be found. These seedlings are dispersed at distances very far from the large plants and, since they do not depend on cross-pollination only for growth but can grow and spread from dislocated vegetative parts and seeds dispersed by birds and other animals, then they are at a great advantage (Hernández-López et al. 2008; Malda & Jimenez 1987; Chuk 2010). At the Prince Albert site, we saw the evidence of spread by vegetative growth in some of the seedlings we measured.

The indication that M. geometrizans is spreading in the study sites is also corroborated by the size frequency histograms. Reverse J-shaped histograms, which we have here, are characterized by a high number of seedlings (Sutton et al. 2018) and indicate self-regenerating, spreading/expanding populations (Sutton et al. 2018; Hett & Loucks 1976). Such populations are expected to persist and continue proliferating should no intervention be made (Sutton et al. 2018).

The Weed Risk Assessment (Gordon et al. 2010; Pheloung et al. 1999) gave us high total, agricultural and environmental scores, thus indicating a high potential risk of M. geometrizans to the agriculture, natural environmental processes and ecology of our three study sites and possibly the other infested sites in South Africa (Riddin et al. 2016). Myrtillocactus geometrizans was listed as a Category 1a alien invasive species in the 2016 National list of Invasive Terrestrial and Fresh-water Plant Species for the NEMBA Act 10 of 2004. These are species which “may not be owned, imported into South Africa, grown, moved, sold, given as a gift or dumped in a waterway.” It is compulsory to control such species and they should be eradicated where possible. Our study shows that M. geometrizans is a high-risk species and thus should be kept as a Category 1a species in the alien invasive species list.

The Tetrazolium seed viability assay results are accurate and close to the results we would have obtained if we would have tested seed germination by planting them (Patil & Dadlani 1993). Therefore, the considerably high percentage, 68.56%, we obtained for M. geometrizans seed viability means that these plants have a high chance of seed germination when fruit are eaten by birds and other animals and seeds are thus dispersed (Reyes et al. 2003). A factor which increases the chances of fruit—and thus seed dispersal—is that M. geometrizans produces a large number of fruit per stem/branch. These are edible to birds, mammals and human beings (Granados & Hernandez 1995). It is also important to note that both smooth and rough soil crusts are suitable for, and positively affect, the seed germination of M. geometrizans (Rivera-Aguilar et al. 2005). These facts make more important and meaningful the high seed viability result.

Though we were not able to obtain quantifiable results in time for the biological control test of this study we do think, however, that Hypogeococcus sp. would still be a good biological control agent candidate for M. geometrizans since it—and its close relative: H. pungens—have been able to successfully control various columnar and candelabriform cacti such as C. jamacaru, H. martinii and Pilosocereus royenii (Sutton et al. 2018; McFayden and Tomley 1984; Carrera-Martínez et al. 2015). This claim would, however, need further testing at the individual plant level for assessing establishment of the biological control agent and testing at the population level to determine if this agent can be regarded as a successful one (Sutton et al. 2018; Carson et al. 2008).

Though M. geometrizans has been shown to be a high-risk species with spreading and proliferating populations and at an advantage for establishment due to its abilities to grow from dispersed seeds and vegetative parts we think it is still possible, at this stage, to eradicate the its infestations. This is because even the highly-infested sites—the ones surveyed in this study—though they are growing, still have low numbers of trees. Individuals in the infestations can still be counted. A process of eradication or control, however, would need to be implemented promptly, without delay, since evidence shows that these populations have a high potential to increase at a significant rate.

ACKNOWLEDGEMENTS The authors extend their gratitude to South African National Biodiversity Institute (SANBI) for funding this project; the Department of Zoology and Entomology and its staff members for providing their facilities and equipment for the carrying out of the work of the project; P. Muskett, I.D. Paterson, K. Jama, A. Ndlovu, K. Weaver, M. Nethavhani and S. Manzana for providing transport to and from the field and their help with the field work, reporting and data analysis; the owners of 10 Adrian van Jaarsveld Street, Adondorp, Graaff-Reinet and Damsedrif farm, Baviaanskloof; the Prince Albert Municipality for availing their land for the cactus survey; S. and R. Dean for providing data and information about the Prince Albert infestation site; and N.M. Gama for her assistance in lab work and data collection.

REFERENCES AGREN, J. & ZACKRISSON, O. 1990. Age and size structure of Pinus sylvestris populations on mires in central and northern Sweden. Journal of Ecology 78: 1049–1062.

BISWAS, T.K., PANDIT, S. & CHAKRABARTI, S. 2009. Tests on Indian and Peruvian medicinal plants used for wound healing. In: Mukherjee, P.K. & Houghton, P.J. (Eds) Evaluation of Herbal Medicinal Products. 228-242. Pharmaceutical Press, London.

BRAUN, P., MACHADO, M. & TAYLOR, N.P. 2017. Cereus jamacaru (amended version of 2013 assessment). The IUCN Red List of Threatened Species 2017: e.T152911A121494380. http://dx.doi.org/10.2305/IUCN.UK.2017-3.RLTS.T152911A121494380.en. Accessed: 03/09/2018.

CARRERA-MARTÍNEZ, R., APONTE-DÍAZ, L., RUIZ-AROCHO, J. & JENKINS, D.A. 2015. Symptomatology of Infestation by Hypogeococcus pungens: Contrasts between Host Species. Haseltonia 21: 14–18.

CARSON, W.P., HOVICK, S.M., BAUMERT, A.J., BUNKER, D.E., & PENDERGAST, T.H. 2008. Evaluating the post-release efficacy of invasive plant biocontrol by insects: a comprehensive approach. Arthropod Plant Interact 2:77–86.

CHUK, M. 2010. Invasive cacti – a threat to the rangelands of Australia. In: D.J. Eldridge & C. Waters Bourke (Eds) Proceedings of the 16th Biennial Conference of the Australian Rangeland Society. Australian Rangeland Society, Perth.

DURÁN, R., TAPIA, J.L. & HERNÁNDEZ, H.M. 2017. Hylocereus undatus (amended version of 2013 assessment). The IUCN Red List of Threatened Species 2017: e.T152183A121525667. http://dx.doi.org/10.2305/IUCN.UK.2017-3.RLTS.T152183A121525667.en. Accessed: 03/09/2018.

GOEDEN, R., FLESCHNER, C. & RICKER, D. 1967. Biological control of prickly pear cacti on Santa Cruz Island, California. Hilgardia 38(16): 579-606.

GORDON, D.R., ONDERDONK, D.A., FOX, A.M. & STOCKER, R.K. 2008. Consistent accuracy of the Australian weed risk assessment system across varied geographies. Diversity and Distributions 14: 234–242.

GRANADOS, S.D. & HERNANDEZ, H.J. 1995. The gathering system in a community of the valley of the Mezquital, Hidaldo. Revista Chipango Serie Ciencias Forestales y del Ambiente 1: 109–115.

HERNÁNDEZ-LÓPEZ, D., VAILLANT, F., REYNOSO-CAMACHO, R. & GUZMAN-MALDONA, S. 2008. Myrtillocactus (Cactaceae): botanical, agronomic, physicochemical and chemical characteristics of fruits. Fruits 63(5): 269-276.

HETT, J.M. & LOUCKS, O.L. 1976. Age structure models of balsam fir and eastern hemlock. Journal of Ecology 64: 1029–1044.

KAPLAN, H., WILSON, J.R.U., KLEIN, H., HENDERSON, L., ZIMMERMANN, H.G., MANYAMA, P., IVEY, P., RICHARDSON, D.M., NOVOA, A. 2017. A proposed national strategic framework for the management of Cactaceae in South Africa. Bothalia 47: 119-130.

KLEIN, H. 1999. Biological control of three cactaceous weeds, Pereskia aculeata Miller, Harrisia martini (Labouret) Britton and Cereus jamacaru De Candolle in South Africa. African Entomology Memoir 1: 3–14.

KLEIN, H. 2002a. Harrisia cactus/ Moon cactus (Harissia martinii). PRI Leaflet Series: Weeds Biocontrol 2.8: 1-2.

KLEIN, H. 2002b. Harrisia cactus mealybug (Hypogeococcus pungens). PRI Leaflet Series: Weeds Biocontrol 2.5: 1-4.

MALDA, G., & JIMENEZ, J. 1987. Commercial production of endangered cacti in arid zones. General Technical Report Rocky Mountain Forest Range Expansion Station, USDA Forest Services Fort Collins, CO, USA, 78–81.

MCFADYEN, R.E. & TOMLEY, A.J. 1981a. Biological control of Harrisia cactus, Eriocereus martinii, in Queensland by the mealybug Hypogeococcus festerianus, In: Del Fosse, E.S. (Ed.) Proceedings of the Fifth International Symposium on the Biological Control of Weeds, 1980, at Brisbane, Australia. 589-594.

MCFADYEN, R.E. & TOMLEY, A.J. 1981b. The successful biological control of Harrisia cactus (Eriocereus martinii) in Queensland. In: Wilson, B.J., & Swarbrick, J.T. (Eds) Proceedings of the Sixth Australian Weeds Conference, 1981, at Gold Coast, Australia. 139-144.

MORAN, V.C., HOFFMANN, J.H. & ZIMMERMANN, H.G. 2013. 100 years of biological control of invasive alien plants in South Africa: History, practice and achievements. South African Journal of Science 109(9/10): 1-6.

MORAN, V.C. & ZIMMERMANN, H.G. 1991. Biological control of cactus weeds of minor importance in South Africa. Agriculture, Ecosystems and Environment 37: 37-55.

NOVOA, A., KAPLAN, H., KUMSCHICK, S., WILSON, J.R.U. & RICHARDSON, D.M. 2015a. Soft Touch or Heavy Hand? Legislative Approaches for Preventing Invasions: Insights from Cacti in South Africa. Invasive Plant Science and Management 8: 307–316.

NOVOA, A., LE ROUX, J.J., RICHARDSON, D.M., & WILSON, J.R.U. 2017. Does commercial horticultural trade in Cactaceae pose a major environmental threat? Conservation Biology 31(5): 1066–1075.

NOVOA, A., LE ROUX, J.J., ROBERTSON, M.P., WILSON, J.R.U. & RICHARDSON, D.M. 2015b. Introduced and invasive cactus species: a global review. AoB PLANTS 7: 1-32.

OLCKERS, T. 2004. Targeting emerging weeds for biological control in South Africa: the benefits of halting the spread of alien plants at an early stage of their invasion. South African Journal of Science 100: 64-68.

PATERSON, I.D., HOFFMANN, J.H., KLEIN, H., MATHENGE, C.W., NESER, S. & ZIMMERMANN, H.G. 2011. Biological control of Cactaceae in South Africa. African Entomology 19(2): 230–246.

PATIL, V. N. & DADLANI, M. 1993. Tetrazolium test for seed viability and vigour. In: Agarwal, P.K. (Ed.) Handbook of Seed Testing. 209–241. National Seed Corporation, New Delhi, India.

PAYNTER, Q. 2005. Evaluating the impact of a biological control agent Carmenta mimosa on the woody wetland weed Mimosa pigra in Australia. Journal of Applied Ecology 42: 1054–1062.

PEREZ, G.R.M., VARGAS, S.R. & ORTIZ, H.Y.D. 2005. Wound healing properties of Hylocereus undatus on diabetic rats. Phytotherapy Research 19: 665–668.

PHELOUNG, P.C., WILLIAMS, P.A. & HALLOY, S.R. 1999. A weed risk assessment model for use as a biosecurity tool evaluating plant introductions. Journal of Environmental Management 57: 239–251.

PORTER, R., DURRELL, M. & ROMM, H. 1947. The use of 2, 3, 5-triphenyltetrazoliumchloride as a measure of seed germinability. Plant Physiology 22(2): 149-159.

REYES, S.J., CASTRO, G.R., MERA, B.S. & BRACHET, I.C. 2003. A note on the establishment of a rural nursery for cacti and succulent production in the Valley del Mezquital, Hidalgo. Cactus Succulents of Mexico 48: 118–126.

RIDDIN, T., VAN WYK, E. & ADAMS, J. 2016. The rise and fall of an invasive estuarine grass. South African Journal of Botany 107: 74–79.

RIVERA-AGUILAR, V., GODINEZ-ALVAREZ, H., MANUELL-CACHEUX, I. & RODRIGUEZ- ZARAGOZA, S. 2005. Physical effects of biological soil crusts on seed germination of two desert plants under laboratory conditions. Journal of Arid Environments 63: 344–352.

SEGARRA-CARMONA, A.E., RAMÍREZ-LLUCH, A., CABRERA-ASENCIO, I. & JIMÉNEZ-LÓPEZ, A.N. 2010. First Report of a New Invasive Mealybug, The Harrisia Cactus Mealybug Hypogeococcus Pungens (Hemiptera: Pseudococcidae). The Journal of Agriculture of the University of Puerto 94(1-2): 183-187.

SIMBERLOFF, D., MARTIN, J., GENOVESI, P., MARIS, V., WARDLE, D.A., ARONSON, J., COURCHAMP, F., GALIL, B., GARCIA-BERTHOU, E., PASCAL, M., PYSEK, P., SOUSA, R., TABACCHI, E. & VILA, M. 2013. Impacts of biological invasions: what's what and the way forward. Trends in Ecology & Evolution 28(1): 58-66.

SUTTON, G.F., KLEIN, H. & PATERSON, I.D. 2018. Evaluating the efficacy of Hypogeococcus sp. as a biological control agent of the cactaceous weed Cereus jamacaru in South Africa. BioControl 63(4): 493–503.

THE STATE OF QUEENSLAND. 2017. Harrisia cactus Moonlight cactus Harrisia martinii, Harrisia tortuosa and Harrisia pomanensis. Biosecurity Queensland 1-4.

TOMLEY, A.J. & MCFADYEN, R.E. 1984. Preliminary indications of success in the biological control of Harrisia cactus (Eriocereus martinii Lab.) in Queensland. Proceedings of the first conference of the Council of Australian Weed Science Societies 108-112.

VAN WILGEN, B.W., DE WIT, M.P., ANDERSON, H.J., LE MAITRE, D.C., KOTZE, I.M., NDALA, S., BROWN, B. & RAPHOLO, M.B. 2004. Costs and benefits of biological control of invasive alien plants: case studies from South Africa. South African Journal of Science 100: 113-122.

VERMA, P. & MAJEE, M. 2013. Seed Germination and Viability Test in Tetrazolium (TZ) Assay. Bio- protocol 3(17): 1-4.

WALTERS, M., FIGUEIREDO, E., CROUCH, N.R., WINTER, P.J.D, SMITH, G.F., ZIMMERMANN, H.G. & MASHOPE, B.K. 2011. Naturalised and invasive succulents of southern Africa. Abc Taxa 11.

WHARTON, M. J. 1955. The use of tetrazolium test for determining the viability of seeds of the genus Brassica. Proceedings of the International Seed Testing Association 20: 81-88.

ZACHARIADES, C., PATERSON, I.D., STRATHIE, L.W., HILL, M.P. & VAN WILGEN, B.W. 2017. Assessing the status of biological control as a management tool for suppression of invasive alien plants in South Africa. Bothalia 47(2): 1-19.

FIGURES AND TABLES

Study sites

Fig.1. Map showing the three study sites.

Population dynamics Population size 90

80

70

60

50

40

30 Number plantsof Number 20

10

0 Baviaanskloof Graaff-Reinet Prince Albert Site

Fig. 2. The population sizes of M. geometrizans in each site surveyed. The total population size is 181 trees.

Table 1. The total number of plants, average size and number of mature plants in each site.

Site Total number of plants Average size (m2) Number of adults Baviaanskloof 71 108.51 14 Graaff-Reinet 56 185.33 23 Prince Albert 54 243.86 40

Distance vs. Size Correlation for Baviaanskloof site Correlation: r = -.1061 2000

1800 Plant 1

1600

1400

1200

) 2 1000

800

Size (m

600

400

200

0

-200 -20 0 20 40 60 80 100 120 140 160 180 200 220

Distance (m) 0.95 Conf.Int. Fig. 3. Correlation between plant size and distance from the oldest cactus for the Baviaanskloof population. The Product-Moment and Partial correlation methods reported a p-value = 0.378383241 (N=71). This means that there is no real significance in this correlation of this infestation.

Distance vs. Size for Graaff-Reinet site Correlation: r = -.3343 2200 Plant 1

2000

1800

1600

1400

1200

)

2 1000

Size (m 800

600

400

200

0

-200 -20 0 20 40 60 80 100 120 140 160

Distance (m) 0.95 Conf.Int. Fig. 4. Correlation between plant size and distance from the oldest cactus for the Graaff-Reinet population. The Product-Moment and Partial correlation methods reported a p-value = 0.0117887369 (N=56). Since this is less than 0.05, a real significance in this correlation of this infestation can be found.

Distance vs. Size Correlation for Prince Albert site Correlation: r = -.2305 700 Plant 1

600

500

400

)

2 300

Size (m 200

100

0

-100 -20 0 20 40 60 80 100 120 140 160 Distance (m) 0.95 Conf.Int. Fig. 5. Correlation between plant size and distance from the oldest cactus for the Prince Albert population. The Product-Moment and Partial correlation analyses reported a p-value = 0.0936440418 (N= 54). There is no real significance in this correlation.

Categorized Histogram of number of branches per site 70 60 50 40 30 20 10 0 -50 0 50 100 150 200 250 300 350 400 450 -50 0 50 100 150 200 250 300 350 400 450

Site: Bavianskloof Site: Graaff Reinet 70 60 50

Number of observations Number 40 30 20 10 0 -50 0 50 100 150 200 250 300 350 400 450

Site: Prince Albert Number of branches

Fig. 6. Plant distribution by number of branches categories in the three sites. One-way ANOVA by ranks and Kruskal-Wallis ((2, N= 181) =27.18674 p = 0.0000).

Table 2. Smallest size when first reproductivity occurs. Site Number Reproductive Size at which reproductive (m2) Number of branches

Baviaanskloof 6 647.4 210 Graaff-Reinet 11 104 26

Prince Albert 28 118 34

Risk Assessment

Table 3. Scores and outcome of risk assessment.

Total score 29 Outcome Invasive (Eradicate or Control) Agricultural score 24 Environmental score 21

Table 4. Seed viability test results. Fruit number Number of seeds Seeds tested positive Percentage positive 1 66 41 62.12 2 47 37 78.72 3 129 68 52.71 4 93 84 90.32 5 53 36 67.92 Total 388 266 68.56

Biological Control Table 5. Initial measurements of planted cacti parameters Species M. geometrizans H. martinii C. jamacaru H. balansae Hylocereus undata Mean number of branches 2.4 2.1 1.4 1.4 1.9 Mean plant height (cm) 32.44 46.91 42.82 46.3 49.28 Mean base of trunk 20.2 10.22 17.76 13.93 8.35 circumference (cm) Mean thickest branch 22 10.22 19.5 8.02 12.62 circumference (cm) Mean diameter of canopy 12.75 30.45 8.7 10.8 17.5 (cm)

APPENDIX 1: Data Sheet

Site: Date:

Collectors:

Plant Plant height Diameter x 2 Number of Reproductive? (fruit, Distance from plant 1 number stems/branches flowers, buds)

APPENDIX 2a: Weed Risk Assessment

Botanical name: Outcome: Common name: Score: Family name: Your name:

History/Biogeography

A 1 Domestication/cultivation 1.01 Is the species highly domesticated? If answer is ‘no’ got to 2.01 C 1.02 Has the species become naturalised where C grown? 1.03 Does the species have weedy races?

2 Climate and Distribution 2.01 Species suited to South African climates (0- low; 1-intermediate; 2-high)

2.02 Quality of climate match data (0-low; 1-

intermediate; 2-high) C 2.03 Broad climate suitability (environmental versatility)

C 2.04 Native or naturalised in regions with extended dry periods

2.05 Does the species have a history of repeated introductions outside its natural range?

C 3 Weed elsewhere 3.01 Naturalised beyond native range

E 3.02 Garden/amenity/disturbance weed

A 3.03 Weed of agriculture/horticulture/forestry

E 3.04 Environmental weed

3.05 Congeneric weed

Biology/Ecology

A 1 Undesirable traits 4.01 Produces spines, thorns or burrs

C 4.02 Allelopathic

C 4.03 Parasitic

A 4.04 Unpalatable to grazing animals

4.05 Toxic to animals

4.06 Host for recognised pests and pathogens C 4.07 Causes allergies or is otherwise toxic to C humans C 4.08 Creates a fire hazard in natural ecosystems E 4.09 Is a shade tolerant plant at some stage of its E life cycle

E 4.10 Grows on infertile soils

E 4.11 Climbing or smothering growth habit

E 4.12 Forms dense thickets

E 5 Plant type 5.01 Aquatic

C 5.02 Grass

E 5.03 Nitrogen fixing woody plant

C 5.04 Geophyte

C 6 Reproduction 6.01 Evidence of substantial reproductive failure in invasive habitat C

6.02 Produces viable seed. C 6.03 Hybridises naturally C C 6.04 Self-fertilisation

C 6.05 Requires specialist pollinators

C 6.06 Reproduction by vegetative propagation C 6.07 Minimum generative time (years)

A 7 Dispersal mechanisms 7.01 Propagules likely to be dispersed unintentionally C

7.02 Propagules dispersed intentionally by people A 7.03 Propagules likely to disperse as a produce C contaminant E 7.04 Propagules adapted to wind dispersal

7.05 Propagules buoyant E 7.06 Propagules bird dispersed

7.07 Propagules dispersed by other animals C (externally)

7.08 Propagules dispersed by other animals C (internally)

C 8 Persistence attributes 8.01 Prolific seed production

A 8.02 Evidence that a persistent propagule bank is formed (>1 yr.) A 8.03 Well controlled herbicides

C 8.04 Tolerates or benefits from mutilation, cultivar or fire E 8.05 Effective natural enemies present in enemies

A = agricultural E= environmental, C= combined

APPENDIX 2b Weed Risk Assessment Scoring Sheet