Plant Hardiness
[Further notes given at the 2001 International Magnolia Society Annual Meeting] © Karen Foley 2026
In March 2000, Jim Gardner asked me to give a lecture at the meeting of the International Magnolia Society in Dublin the following year. He suggested that I should speak on the Climatic Zone 9 plants that I grow in my garden at Earlscliffe, Baily, Co. Dublin at 53° north latitude. On learning during a second telephone conversation in February 2001 that my garden was experiencing one of the coldest winters since I started taking records 31 years ago and that many plants had been killed, he suggested that I speak instead on ‘Plant hardiness’.
So I want to look at the hardiness of Climatic Zone 9 plants in Ireland and at low temperature damage generally. I imagine that most people will know that the idea of climatic hardiness zones was first developed for the US and has been extended to the rest of the world. Climatic zones 1 to 9 are based on the average of the absolute minimum temperature over a 20-year period. Climatic zones give an indication of how hardy a plant is. They are not perfect, but they are a good place to begin. The whole subject is extremely complex, as many other factors are involved in plant survival.
The main factor influencing plant hardiness is the absolute minimum temperature, but this is only one of many factors. Other factors include the duration of freezing temperatures, the speed of drop, the hardening-up process, the speed of thaw, the moisture content of the plant, the soil and the air, and soil type. Healthy, well-grown plants will be better able to cope with cold than unhealthy plants, so a wide range of cultural factors also come into play, including nutrition, plant health and general husbandry. Plant morphology also matters, with some plants having especially susceptible stems, buds or roots. In addition, there are other climatic influences such as wind chill and depth of snow cover.
There is still much argument among scientists about the physiological basis of cold injury. Freezing injury can be caused by intracellular freezing of the cell contents or indirectly by dehydration of tissues resulting from extracellular freezing. Plants are usually killed by formation of ice crystals within the cells, but formation of ice crystals in the intercellular spaces is not necessarily fatal. If frost-hardened plants are cooled slowly, ice initially forms in intercellular spaces and, as the temperature decreases, water moves out of the cells to these ice nuclei. As a result, the concentration of the cell sap is increased and it was once thought that this accounted for the lowering of the freezing point. It is now believed that this effect is small, and it now seems that, in the more hardy species, plant tissue avoids freezing by deep supercooling to temperatures of −20 °C or less. Other workers have placed emphasis on the structural nature of protein molecules when water is withdrawn from the cell during the formation of intercellular ice crystals. It has been suggested that the accumulation of solutes within the cell at low temperatures may prevent the denaturing and aggregation of proteins by osmotic retention of water. This applies not only to proteins within the cell but also to those incorporated in the surrounding protoplasmic membrane. If proteins associated with the membrane were irreversibly damaged, there would be loss of semipermeability and death of the cell.
If frost-hardened plants are cooled rapidly they are often injured, possibly because water does not move out of the cells quickly enough and intracellular freezing suddenly occurs. Killing by intracellular freezing may be due to mechanical disruption of the protoplasm or alternatively to dehydration.
Freezing injury can take many forms, such as killing back of shoots, formation of frost rings, injury to stems or roots, and frost heaving.
Shoot injury
Frost can damage partially hardened shoots or fronds in tree ferns. Late growth is not unusual in a temperate oceanic climate such as Ireland’s, and may be exacerbated in some countries by late application of fertiliser or irrigation water. Such late growth does not have time to develop cold hardiness. Sudden killing of leaves prevents the formation of an abscission layer and such leaves often remain attached all winter. While it is not possible to generalise, plants that form an abscission layer and drop their leaves cleanly appear to have a better chance of recovery than those where no abscission layer forms.
Nicking the stem of plants with a fingernail a month or so after the cold spell gives a good indication of the chances of survival. If the inner bark is green and the cambial cells and young xylem are moist, there is a good chance of survival even if the plant has lost all its leaves. But if the inner bark, cambium and xylem are brown, hard and dry, then the prognosis is not good.
Stem injury
Frost cracks and winter sunscald lesions sometimes develop on stems of thin-barked trees when alternating freezing and thawing occur during the winter. Lesions can be recognised as dead bark patches, bark peeling back to reveal wood, or sunken cankers that become sites for fungal attack. These lesions damage the phloem so that the downward translocation of organic solutes is impeded. Bark on the south side may be 20 °C warmer at midday than bark on the north side, suggesting that the south side undergoes violent temperature fluctuations. Heating of the south side may initiate changes that decrease the cold hardiness of the cambium and make it very susceptible to frost injury.
Root injury
Stem tissues are hardier than root tissue, but roots are injured less in winter because soil and snow cover protect them to some extent from exposure to freezing temperatures. Nevertheless, when soil freezes, roots are often damaged, especially the small, physiologically important ones. The location of roots in the soil is important both for plant health and for freezing injury.
As soil management in my garden is virtually entirely by means of herbicides and the soil surface is not disturbed, plant roots are close to the surface. Many self-seeding plants, such as Solanum laciniatum, never put roots deeper than 2 cm, so the entire root system is right at the soil surface. This would suggest that root injury would be especially serious in my garden, as the surface soil at least was frozen during the severe cold spell, but in general this has not been the case. A large number of roots were examined about seven weeks after the severe cold spell and, as far as could be ascertained, root damage was minimal compared with the severe injury to the top parts of the plants. In some cases it was clear that both top and root were damaged, for example Euryops chrysanthemoides and Aeonium, but in many cases, although the top was seriously affected, the roots appeared turgid, moist and fresh. Plants in this latter category were Solanum laciniatum, Cyathea cooperi, Psoralea affinis, P. pinnata, Tetrapanax papyfera, and Eucalyptus ficifolia.
It is clear from the way so many plants resprout strongly from the base, perhaps six months or one year after the top growth has been entirely killed, that in many plants the Achilles heel is not the roots as far as low-temperature injury is concerned. It has been suggested that roots develop frost resistance only to temperatures ranging from about −2 to −10 °C.
Frost hardiness of roots varies with species and rootstock. Cold injury to roots appears to be worse in sandy than in clay soils.
Different types of injury
I have found it fascinating wandering round the garden seeing the many different ways that plants show low-temperature injury.
Some show damage immediately after the first severe frost, say −4 or −5 °C. Such plants include Aeonium tabuliforme and a number of other succulents. These plants grow in low-rainfall areas and contain exceptionally large quantities of water. The appearance of a water-soaked condition the morning after a cold night suggests that cell structure has broken down, either because of damage to the cell walls or to the semipermeable membrane surrounding the cells. Within a short time it is hard to recognise the amorphous mass as a plant at all.
Another plant that goes quickly at about −5 °C is Brugmansia sanguineum. Where the temperature does not fall below −5 °C, stems remain intact although flowers and leaves will be damaged, and the plant can grow to four or five metres. Although this year the stem looks damaged, if previous experience can be relied on the plant will sprout from the base.
In contrast to these plants where damage is rapid, Eucalyptus ficifolia can look fine for some weeks after frost. The shoot tips, which remain active during a normal winter, will be killed early on, but after a severe frost, the leaves may remain apparently in good condition for weeks before eventually turning a deep brown colour. The cambium, phloem and xylem of the stem structure appear killed, but the roots seven weeks afterwards were alive.
Sonchus arboreus was affected in an unusual way. The tips of branches were completely killed and dry. Lower down, the bark, phloem and cambium appeared mushy, but the roots appeared in good condition.
During the growing season plants have little ability to withstand frost, but killing frosts do not usually occur during this period. With the approach of autumn, changes occur in the plant that give it a degree of hardiness, with further increases during winter. The problem with many Zone 9 plants is that they do not recognise the signs that winter is approaching, such as shortening autumn days. They do not stop growing and do not start acquiring any real degree of hardiness.
Plants vary in their degree of hardiness during the winter. Temperature fluctuations during late winter can also bring about changes in hardiness. A mild spell in early February can result in a decrease in hardiness. It is not possible, therefore, to describe any plant only in terms of the lowest temperature that it will withstand. Other factors include:
- Time of onset of hardening
- Rate of hardening
- Absolute degree of hardiness attained
- Ability to recover hardiness after a mild spell
- Time of onset of dehardening
In Ireland’s mild climate it is these erratic winter temperature changes that are potentially damaging.
Frost damage itself is still debated. It is agreed that water is present in plant tissue in two main fractions: in the cell wall and within the cell itself in the protoplast and vacuole. Ice forms first in the cell walls and water moves out of the protoplasm, through the cell membrane, and adds to this accumulation. As the temperature falls further, death point is reached. This is where the argument starts. One group believes that ice crystals form in the protoplast and disrupt function, while the other believes that continuing movement of water out of the protoplast induces dehydration that disrupts it.
Some plants develop considerable tolerance. Physiological changes take place during the period we think of as dormancy. The net result of these changes determines the extent to which plants can cope. So it is not possible to look back at records and say what plant will stand up to what level of frost. There are too many permutations of mild and cold spells for confident predictions.
Efforts to increase plant hardiness
We cannot control the weather, but we can do something to increase plant hardiness. For instance, much could be done to increase the hardiness of Climatic Zone 9 plants to Irish conditions by selection and breeding.
For example, Eucalyptus ficifolia flowered well at Earlscliffe between September and November 1976, but trees were killed by frost during the 1976–77 winter. This species does not usually survive temperatures lower than −1 to −4 °C. The brilliant red blossom and attractive leaves with a reddish mid-rib and margin made the flowering plant in 1976 the most attractive tree I have ever seen. The memory of it encouraged me to attempt to find a hardier clone at the first opportunity (see * below).
This did not arrive until January/February 1996 when my wife and I set off for Western Australia specifically to collect seed of E. ficifolia in the ‘coldest’ area possible and to learn something about hybridising Eucalyptus at the University of Western Australia, Nedlands, Perth.
Having obtained the necessary licence to collect seed for scientific purposes from the Department of Conservation and Land Management (CALM) in Perth, we set off for the south coast. Unlike many eucalyptus species, E. ficifolia has a very restricted range and occurs naturally in only two small areas: inland about 30 miles north-east of Walpole near Mount Frankland, and right on the coast about nine miles east of Walpole. The Mount Frankland area is under 300 metres above sea level, so there may not be a great difference in climate harshness between the two areas. Seed was also collected from trees planted in other colder parts of Western Australia.
After travelling about 11,000 miles by air and 300 miles by car, it was an exciting moment spotting the first E. ficifolia tree, about eight metres tall, with a dome of brilliant red flowers and a good crop of seed capsules from the previous year’s flowering. Unfortunately this first tree had an unbranched trunk to about three metres, making it impossible to climb, so we travelled further in search of more accessible seed. Capsules were collected from both wild areas, and the large seed was rapidly extracted by allowing the capsules to dry in a glasshouse. Ambient temperatures were over 40 °C most days.
I envisaged that if I planted large numbers of seedlings closely together at Earlscliffe, the first winter would kill most of them (seedlings are usually most susceptible in their first season) and, with some luck, leave one or two hardier plants that could be multiplied by tissue culture.
Seed was sown at Earlscliffe in cold glass in April 1996 and germinated readily. Some seedlings were planted out in autumn, and the majority in spring 1997. Paradoxically, none of the seedlings outdoors has been killed (1995–96 was the last cold winter), but just under 50% of seedlings overwintered under glass in the first season were lost, probably due to Phytophthora. By December 2000, I had over 100 young plants, some over three metres tall, all planted too closely together (about 18 cm apart). None had yet flowered, but all looked in good condition. The frosts of late December 2000 and February/March took their toll, but one or two plants may have survived.
Plants have been distributed to other locations including Fota and Malahide Castle.
* I am also attempting to hybridise, but ficifolia will not cross with gunnii or pulverulenta. Red-flowered caesia will, and I have obtained pollen from Western Australia and done some crossing, but the pollen was weak on arrival. I have a tree of caesia at Glasnevin and I hope to obtain fresh pollen from it as soon as it flowers and continue.
Echium
The Giant Borage is obviously not native to Co. Dublin, yet this plant, Echium pininana, has made itself completely at home here. With its enormous, triffid-like spire, unusual leaf form and unbranching structure, it makes a powerful statement in any landscape or garden.
The Giant Borage comes from the Canary Islands, yet paradoxically it is highly endangered in its native land. Like foxglove, it is monocarpic, meaning it flowers and seeds only once (from mono meaning once, and carpel, a seed). Whenever the plant decides to flower, whether in its first, second or third year, every bit of energy goes into producing the flowering spike. By year’s end the plant is depleted and dies. Even in death, there is a certain sad beauty about this dignified corpse.
In its native land it grows on harsh volcanic slopes, so it is not used to an affluent lifestyle. A good-sized borage can produce one million seeds and the ground can be carpeted with seedlings. You do not even have to thin them out: the most vigorous kill the weaker and the plants self-select.
In my early gardening years at Earlscliffe, cold winters such as 1978–79 and 1986–87 killed all Echium seedlings. The harder winters of 1995–96 and 2000–01 have not. It appears that a hardier form is evolving. This is not surprising, as in some winters a proportion of seedlings will die, so seed is shed mainly from survivors.
Here is a related species, the red borage, Echium wilpretii, which is even more attractive but less vigorous and does not self-seed. I collected seed from one parent plant in late 1998 and sowed it in spring 1999. The plant forms a woolly ball in the first year and flowers in the second. An interloper arose from seed from the same parent. A bee presumably carried pollen from the Giant Borage to the Red Borage, producing a hybrid. Its hybrid nature is obvious: it is intermediate in height and flower density, and its carmine/purplish colour mixes the blue flowers of its father with the red of its mother.
Between April and mid-June the Giant Borage grows at about 2 inches (50 cm) a day. Then, on the longest day, its upward surge stops abruptly and it gets on with dispersing its seed.
The hybrid E. wilpretii × E. pininana appears hardier than either parent, as it does not shrivel in the cold as much as E. pininana.
E. nervosum appears very sensitive to cold, as two plants were killed last winter. E. candicans also appears susceptible and is looking very sick at present.
Conclusions
Given the diversity of plant responses to low temperature and the diversity of Irish weather patterns, it is not surprising that it is difficult to identify the precise reasons for seemingly erratic winter injury to tender plants.
Mike Nelhams, Curator of Tresco Gardens on the Isles of Scilly, tells me that Embothrium will not grow with him, which suggests that some other factor is involved, as Embothrium is very hardy, can withstand −10 °C and is Climatic Zone 8.
The highly endangered palm, Juania australis, is reputed to have failed on Tresco, yet it appears exceptionally hardy in my garden. So the only way to find out if a plant is hardy in a particular area is to plant it and observe its performance over the next 50 years.
Where temperature drops rapidly to a very low level, you can have direct injury where the cell contents freeze before water can move out into the intercellular space. In this case, death of plant tissue is rapid. Or you can have indirect damage where so much water moves out that cell contents dehydrate. In this case, death can be prolonged.
When dealing with almost any aspect of plant growth, we should consider the whole plant, but seldom do. It would be better if we could see the whole plant, as the whole plant is very much involved in hardiness. Because of my research interest in chemical weed control, I have always had a special interest in roots.
Some plants die very slowly, for example Psoralea pinnata. Presumably this is due to dehydration of the cell contents because the dead plant looks as if it has been dehydrated.
It is interesting to observe how plants killed or damaged by low temperatures respond in a wide variety of ways. While scientists continue to argue about the exact way plants are damaged by frost, there is general agreement that cells are killed if ice crystals form within them. But not all plants that collapse quickly during low temperatures are killed, as many recover. Conversely, some plants can look in reasonable condition for over a month after freezing only to die later in the year.
To conclude, the basic mechanism of frost hardiness is not fully understood, but hardening is accompanied by one or more of the following changes:
- Tissue may become capable of limiting the location of extracellular ice formation to where it causes least damage.
- There is often an increase in dissolved substances in the cell sap, frequently sugars or amino acids, which rather than having a simple anti-freeze effect are thought to protect specific sensitive proteins and enzymes from freezing.
- Plant cells may possess a degree of dehydration tolerance, the basis of which is not fully understood, but may include limitation of cell and vacuole size and synthesis of desiccation-resistant proteins during hardening.
Other pages related to David Robinson
Page last updated: 18 Jul 2026